faking it or muddling through? understanding ......tributing to the data collection. we are indebted...

* Academy of Management Joumal2012. Vol. 55, No. 6. 1429-1448.http://cbc.dci.org/10.5465/amj.2010.0697

FAKING IT OR MUDDLING THROUGH? UNDERSTANDINGDECOUPLING IN RESPONSE TO STAKEHOLDER PRESSURES

DONAL CRILLYLondon Business School

MAURIZIO ZOLLOUniversità Bocconi

MORTEN T. HANSENUniversity of California, Berkeley

We advance a multilevel argument that challenges and qualifies existing explanationsof firms' responses to institutional pressures. In an in-depth study of 17 multinationalcorporations involving 359 interviews with internal and external actors, we find thatfirms facing identical pressures decouple policy fi-om practice in different ways and fordifferent reasons. When firms' responses are generated locally, without firmwidecoordination, these responses can be either intentional or emergent. In the presence ofinformation asjrnunetry between firms and their stakeholders, we find that managers'responses are intentional ("faking it") and depend on how they perceive their interests.In the presence of competing stakeholder expectations, responses are emergent ("mud-dling through") and depend on the degree of consensus among managers in theirreadings of the environment. These findings suggest that theories of decoupling need tobe broadened to include the role of "muddling through" and the interplay of intemalmanagerial and external stakeholder dynamics.

Our industry operates in a diverse environment. Theschizcphrenia makes business difficult. Diverseopinions build conflicting values, and we can'tmanage ever)fthing from the center.

-Multinational chemicals company executive

Why do many firms not implement the policiesthat they adopt? Existing theories of decouplingpaint a cjmical picture of corporate life. Macroleveltheories posit that environmental circumstances—for example, confiicting stakeholder expectations—lead firms to adopt inconsistent standards that theycannot implement simultaneously (Meyer &Rowan, 1977). The emphasis in this literature is ondecoupling as a convenient arrangement betweenfirms and stakeholders, who turn a blind eye to the

This research was funded by a grant provided by the6th Framework Program of the European Commission(Directorate-General for Research) for a project entitled"Understanding and Responding to Social Demands onCorporate Responsibility (RESPONSE)." We thankLoiu-des Casanova, Kai Hockerts, Mario Minoja, PeterNeergaaid, Esben Pedersen, Francesco Perrini, SusanSchneider, Pamela Sloan, and Antonio Tencati for con-tributing to the data collection. We are indebted to Asso-ciate Editor Tima Bansal and three anonymous review-ers, whose valuable feedback and suggestions helped toimprove the article significantly.

state of practice inside firms (Meyer & Rowan,1977: 359). Of course, not all firms necessarily re-spond identically to the same environmental pres-sures (Fiss & Zajac, 2006). Recognizing this, strat-egy scholars have used firm-level variables toexplain why some firms decouple policy from prac-tice even when their peers in the same environmentdo not (Westphal & Zajac, 1994). The emphasis inthis literature is on decoupling as a form of calcu-lated deception. For example, firms deceive stake-holders about their plans when powerful CEOsstand to lose from implementing policy (Westphal& Zajac, 2001).

Although the focus on firm-level variables hasbeen an important development in explaining whyfirms respond to stakeholder pressures in differentways, the variables emphasized in current theory,such as the power and interests of CEOs, are mostrelevant when single actors direct firms' responses.Firms, however, are composed of individuals withtheir own perceptions and interests. Their execu-tives do not always act in unison (Pratt & Foreman,2000), calling into question the dominant under-standing of decoupling as a well-coordinated re-sponse. As individual decision makers have tomake sense of the environmental pressures thattheir firms face (Ceorge, Chattopadhyay, Sitkin, &Barden, 2006), there is no guarantee that their per-

1429Copyright of îhe 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 pemnssion. Users may print, download, or email articles for individual use only.

1430 Academy of Management Joumal December

ceptions will converge when external constituen-cies push for different policies (see, e.g., Fiss &Zajac, 2004). Thus, a rohust explanation of decou-pling needs to open up a hlack hox to take intoaccount how hoth the intemal organization andexternal environment of firms interact in shapingtheir responses to stakeholder pressures.

In this study, we seek to answer the question.Why do firms respond differently to the same insti-tutional pressures? In posing this question, we aimto explain variance both in a single context (that is,why some firms decouple whereas their peers im-plement policy) and across different contexts (thatis, whether the factors explaining decoupling differaccording to the external pressures facing a firm).We are thus interested in "causal recipes" (Ragin,2008: 23): how intrafirm and institutional levelforces comhine to shape responses that cannot beadequately explained by factors at any single levelof analysis in isolation.

To explore this question, we designed a fieldstudy to investigate 17 firms, matched in pairs andtriads by industry and geography but differing intheir implementation of corporate social responsi-hility (CSR) practices. To understand external pres-sures, we conducted 190 interviews with stake-holders. To understand managerial perspectives,we conducted 169 interviews with companyexecutives.

Ovir contrihution is a multilevel framework thathighlights that responses to institutional pressuresdepend on hoth the intemal organization and ex-ternal environment. By comhining these levels ofanalysis, we show that decoupling does not neces-sarily involve intent on the part of managers.Rather, decoupling can he the outcome of organi-zational learning efforts that are fraught with com-plexity under conditions of inconsistent, and rap-idly changing, stakeholder pressures.

DIVERGENT RESPONSES TO EXTERNALSTAKEHOLDER PRESSURES

We discuss two characteristics of environmentsthat play a central role in existing theories of de-coupling: how consistent diverse stakeholders arein their expectations about firms (stakeholder con-sensus) and how aware stakeholders are of the stateof practice inside firms (information asymmetry).Neither characteristic hy itself is likely to predictfirms' responses to stakeholder pressures. As weargue below, simultaneous attention to the externalenvironment and firms' internal organizationpromises to provide more robust explanations andto generate new theoretical insight (Thornton &Ocasio, 2008).

Stakebolder Consensus and ManagerialConsensus

In their seminal work, Meyer and Rowan (1977)explained decoupling as a function of inconsistentstakeholder expectations. When diverse stakehold-ers press for mutually incompatible policies, de-coupling allows executives to mitigate conflictwith stakeholders (Ceorge et al., 2006). Competingstakeholder expectations also weaken the pressureson firms to comply with policy and thus increasetheir discretion over how to respond (Purdy & Cray,2009). Evidence confirms that, when unsure oftheir reading of the environment, executives initi-ate responses aimed at accommodating distinct fac-tions in that environment. For example, even afterespousing the Java software technology. Sun Micro-systems sponsored the development of alternativestandards (Carud, Jain, & Kumaraswamy, 2002).

Yet firms do not have to reflect the diversity oftheir environments because understandings of le-gitimacy are also constructed internally (Basu &Palazzo, 2008; Battilana & Dorado, 2010). NovoNordisk, for example, singles out patients as itsultimate stakeholders (Novo Nordisk, 2011), resolv-ing uncertainty for managers confronted by com-peting expectations from patients, health care pro-viders, regulators, suppliers, and communities.Executives can forge common understandings bymeans of formal policies and sanctions. By restrict-ing subordinates' influence over resource alloca-tion, they limit managerial discretion and poten-tially produce a tighter coupling between policyand practice (Camoran & Dreeben, 1986). Even inthe absence of formal controls, use of shared valuesand strong leadership to channel organizational at-tention can unify organizations (Pratt & Foreman,2000). Such unity implies that decoupling is not aninevitable response to inconsistent stakeholderexpectations.

If firms, when faced with competing stakeholderexpectations, do decouple policy from practice,there are also grounds for attending to their intemalorganization. Most accounts depict decoupling asan intentional response by executives. However, aplethora of standards to choose from creates confu-sion for executives, and delegating authority tomiddle managers appears to he a sensible solution(World Bank, 2003). The discretion of managers atvarious levels of an organizational hierarchy maylead to the inconsistent implementation of head-quarters policy. In the ahsence of certainty abouthow best to respond, organizations engage in prob-lem-focused search (Cyert & March, 1963), whichincreases variety within a firm as individual unitsseek distinctive local solutions (Leifer, 1988). Fur-

2012 Crilly, Zoilo, and Hansen 1431

ther, even if units discover solutions that are ap-propriate for elsewhere in their firm, imperfectlearning can impede their replication (Winter &Szulanski, 2001). From this perspective, decou-pling might result from variation within a firmrather than from coordinated decision making atthe top.

Information Asymmetry andOrganizational Interests

Though the seminal studies of decoupling assertthat stakeholders are often complicit in firms' de-coupling and purposely keep their monitoring offirms to a minimum (Meyer & Rowan, 1977: 359),information asymmetry—discrepancy in the infor-mation levels of managers and stakeholders,wherein the managers are better informed abouttheir firms' practices (Akerlof, 1970; Kulkarni,2000)—is central to the more recent strategic ac-counts of decoupling. It can be difficult for stake-holders to observe firms' internal practices (Christ-mann & Taylor, 2001). Under such conditions,executives can deceive stakeholders about the stateof practice inside their firms to gain legitimacywhilst pursuing their personal or organizationalinterests. For example, the contradiction betweenBritish American Tobacco's formal health andsafety policies and the lack of protective clothingworn by its laborers in Brazil and Kenya occursbecause stakeholder access to local information isrestricted (Christian Aid, 2004).

However, this depiction of firms' responding op-portunistically to institutional pressures poten-tially underplays their need for ongoing, produc-tive relationships with external stakeholders.Firms' self-interest is seldom as unambiguous asthe interest of individual CEOs in maintaining theirown compensation and can go beyond minimizingthe costs of compliance with social expectations.Rather, executives can internalize pressures forcompliance in different ways (Sonenshein, 2006),perceiving these as sources of opportunity or threatfor their firm (George et al., 2006). Threats includerisks to the firm's legitimacy and resources,whereas opportunities include increased access tomaterial resources.

How executives perceive institutional pressuresmatters because these perceptions are likely to in-fiuence the responses that they formulate. Thoughfirms exploit information asymmetry to concealnoncompliance and can sometimes engage in S5mi-bolic behavior without apparent sanction, asymme-try is not an unmitigated blessing (Nayyar, 1990).By impeding stakeholders from identifying trust-worthy corporate partners, asymmetry dissuades

them from collaborating with firms (Axelrod,1984). Decision makers who perceive opportunitiesin compliance with institutional pressures mayhave an interest in reducing asymmetry with theirstakeholders to encourage closer collaboration, po-tentially benefiting the firm (Jones, 1995).

Explanations of responses to institutional pres-sures thus have to take account of firms' require-ments for ongoing relations with their stakeholders.Firms' regular public disclosure may be ineffective(Leland, 1979) because stakeholders may be unableto interpret or trust data released by firms anddiscount their reliability. An alternative approachis to engage in close collaboration with stakehold-ers that allows their participation in, and directobservation of, internal operations (Sivaramakrish-nan, 1994). Close collaboration is likely to involvethe exchange of private information and "new col-laborative opportunities" (Uzzi & Cillespie, 1999:33). An alternative approach is to adopt meaningfulthird-party certification to communicate compli-ance (King, Lenox, & Terlaak, 2005). Thus, the as-sumption that firms will always try to exploit stake-holder ignorance is open to debate.

In summary, dominant theories of decouplinglargely overlook the potentially complex interplaybetween external environment and internal organ-ization. Our contention is that forces at both levelsof analysis combine to shape firms' responses toinstitutional pressures.

EMPIRICAL ANALYSIS

Drawing on our theoretical development, we de-signed an investigation to understand why firms insimilar environments respond differently to insti-tutional pressures. We focused on institutionalpressures for corporate social responsibility (CSR),which we define as a firm's "considerations of, andresponse to, issues beyond the narrow economic,technical, and legal requirements of the firm toaccomplish social benefits" (Davis, 1973: 312).Following the approach of recent scholars, we ex-tend "social" considerations to include environ-mental ones (Aguilera, Rupp, Williams, & Canapa-thi, 2007).

CSR provides a suitable context for understand-ing divergence in responses. Some firms adopt CSRpolicy but decouple it from their activities (Weaver,Treviño, & Cochxan, 1999), diverting resources topublic relations and political efforts (David, Bloom,& Hillman, 2007). Conversely, others comply sub-stantively by adapting their operating processesand integrating social and environmental dimen-sions into their line management (Greening &Gray, 1994).

1432 Academy of Management Journal December

On the basis of an analysis of social rating data,we selected 17 corporations in seven sectors asdefined by type of industrial activity. This mid-sized sample allowed us to combine in-depthknowledge of individual cases and variation acrosscontexts (Cress & Snow, 1996). Our cases are sum-marized in Table 1. We grouped the firms into pairsand triads, in each of which the firms had the sameindustry, size, and level of geographic expansionbut different implementations of policy. Each pairor triad consists of at least one integrator firm,characterized by high substantive action (espousaland implementation of CSR policy), and a decou-pler firm, characterized by high symbolic action(espousal of CSR policy but limited and inconsis-tent implementation).^

Data

We relied on four data sources: (1) social perfor-mance data provided by three rating agencies, (2)

' Tbe firms were also matcbed as to headquarters lo-cation (Nortbem Europe, Southern Europe, Anglo-Saxoncountries), except in the pharmaceutical pair and thehigh-technology triad, in each of which one firm wasfrom Nortbem Europe and its peer or peers were from theAnglo-Saxon regions. If we drop the pharmaceutical pairand tbe Nortbem European high-technology firm fromtbe SEimple, our ñndings remain substantially tbe same,and tbe measures of fit (consistency, coverage) are within5 percent of tbe measures reported for tbe entire sample.

169 interviews with executives in the 17 firms, (3)190 interviews with external stakeholders in sevensectors, and (4) archival data.

Social performance data. To guide our sampleselection, we accessed data from three social rat-ing agencies: Innovest, e-Capital Partners, andVigeo.^ The agencies assess behaviors relevant topolicy adoption (e.g., the visibility of CSR indocuments, and the adoption of environmentaland social standards) as well as substantive ac-tivities (e.g., compliance with standards). Weused 2005 ratings. After our study, we checkedthese for consistency with the ratings valid at theend of 2006. To confirm firms' state of practice,we triangulated by comparing the ratings withinformation derived from executive interviewsand archival evidence.

Executive interviews. Senior managers formu-late responses to issues arising in their sociopo-litical environments and translate environmentalpressures into organizational actions (Kaplan,2008a). From late 2005 to early 2007, we con-ducted 169 semistructured interviews. Our entryto firms was via the executives overseeing CSR.At each firm, we interviewed the CEO or chair-person of the board and the executives responsi-ble for the principal business functions (finance,marketing, human resources). We also inter-

^ KLD ratings were inappropriate as they excludedfirms beadquartered outside of North America.

TABLE 1Description of 17 Cases Studied

Company

SMARTTECHNICSILICONNUCLEUSFUSIONSCIENTIFICEXCAVATIONRESOURCEMINERPOWERENERCETICHIGHFINANCECAPITALEDIBLECONSUMPTIONDRUGMEDICAL

Industry

Higb-tecbHigb-tecbHigb-tecbCbemicalsCbemicalsCbemicalsExtractiveExtractiveExtractiveEnergyEnergyBankingBankingFood & drinkFood & drinkPbarmaceuticalPbarmaceutical

Degree ofImplementation

HigbLow/mediumLowHigbLow/mediumLowHigbLowLowHigbLow/mediumHigbLow/mediumHigbLowHigbLow/medium

Headquarters"

Anglo-SaxonN. EuropeAnglo-SaxonN. EuropeN. EuropeN. EuropeAnglo-SaxonAnglo-SaxonAnglo-SaxonS. EuropeS. EinropeS. EuropeS. EuropeN. EuropeN. EuropeAnglo-SaxonN. Europe

Interviewswith

Executives

1287

10151010

97

12787

1314

713

Interviews withNon-Headquarters

Executives

6242

en

224200004521

Interviewswith

Stakeholders

81016

88

11 .1813

99

121111 .1217

710

" "Anglo-Saxon" refers to tbe United Kingdom and Nortb America. "Nortbem Europe" denotes Scandinavia, tbe Benelux countries, andtbe German-speaking countries. "Southern Europe" refers to Italy and Spain.


viewed regional heads in most firms; the excep-tions were the four domestically focused firms inthe energy and finance sectors. Table 1 providesa breakdown of the interviewees. We conducted110 interviews on site and the remainder by tele-phone. Interviews lasted between 50 and125 minutes (78 minutes on average). Most of theinterviews (141) were recorded and transcribed.Detailed notes were made in the remaininginterviews.

Interviews addressed understandings of CSR andrationales for responding in different ways. Wethen narrowed in on CSR-related practices, includ-ing investment decisions, operations, incentives,and learning.

Stakeholder interviews. We conducted 190 in-terviews, either on-site or by telephone, with stake-holders. Interviews were conducted from late 2005to early 2007 and lasted between 35 and 120 min-utes (60 minutes on average). As with the executiveinterviews, most of the stakeholder interviews(152) ware recorded and transcribed. Detailed noteswere taken in the remaining interviews. We usedthe typology of Post, Preston, and Sachs (2002) tocategorize stakeholders as follows: 58 percent ofour interviewees represented sociopolitical stake-holders (e.g., nongovernmental organizations[NGOs], activist groups, and community entities);18 percent represented resource-based stakehold-ers (e.g., customer associations, and socially re-sponsible investment funds); and 24 percent repre-sented industry structure stakeholders (e.g.,regulators and unions). The focus on sociopoliticalstakeholders and interest groups, rather than spe-cific transactional stakeholders, is appropriategiven their function of monitoring corporate prac-tice (Fedderson & Gilligan, 2001) and influencingpolicy adoption (den Hond & de Bakker, 2007).Managers helped to identify stakeholders by mak-ing lists and informing us whether interactions hadbeen positive or negative. However, the final selec-tion was ours, and we ensured, as far as possible,consistency in terms of numbers and types of stake-holders across firms and sectors.

Our interviews with stakeholders addressed theirunderstanding of CSR in a particular sector andtheir expectations regarding the kinds of CSR activ-ities in which corporations in that sector shouldengage. Each interview was structured around afocal firm. Stakeholders described the history ofinteraction, and we probed for knowledge about thefirm's operating practices.

Documentary evidence. We collected extensivedocumentary evidence, including public reports aswell as internal strategy documents and HR hand-books. We used this evidence to triangulate the

social rating data and to confirm the extent of im-plementation in the 17 firms.

METHODS

We built 17 case histories, combining evidencefrom interviews and archival data to build accountsof CSR-related policy and practice and the ratio-nales that executives used to explain responses.Building on these cases, we turned to fuzzy setqualitative comparative analysis (fsQCA) to iden-tify the conditions associated with implementationand decoupling. The premise of set-theoretic meth-ods such as fsQCA is that causality in the socialsciences is often conjunctural (Ragin, 2008): multi-ple conditions combine to produce outcomes. Thisview of causality is distinct from that in variable-oriented research, which identifies general patternsof association and seeks causes that apply in allcontexts. Using Boolean algebra, fsQCA identifiesthe configurations of conditions associated with anoutcome of interest. Intuitively, we treated a firm asa member of multiple sets (for example, the set offirms with high managerial consensus about CSR)and employed fsQCA techniques to identify consis-tent patterns between set memberships and firms'responses.

This type of analysis had a number of advantagesfor tbe present study. Though some recent fsQCAstudies follow a deductive approach (e.g., Fiss,2011), the method lends itself to the use of smallerdata sets for the purpose of theory elaboration (Red-ding & Viterna, 1999). By identifying how effectscombine to produce outcomes, fsQCA is particu-larly appropriate for advancing multilevel theory(Lacey & Fiss, 2009). Its usage is thus consistentwith our aim of understanding the interplay be-tween factors at the intrafirm and institutional lev-els in shaping responses. Lastly, fsQCA allows forasymmetry (Fiss, 2011) between the drivers of de-coupling and the drivers of implementation. Thisenables a more niianced analysis than conventionalquantitative techniques that would use the sameregression function to explain both implementationand decoupling.

Set membership does not have to be binary (0/1).Rather, in fsQCA the aim is to calibrate set mem-bership in such a way that levels of membershiprepresent meaningful groupings (Ragin, 2008). Lev-els can be 0, 0.33, 0.66, and 1, where 0 representsnonmembership, 1 represents complete member-ship, and 0.33 and 0.66 represent intermediate lev-els; a value of 0.33 implies that a case is more outof, rather than in, the set, and a value of 0.66 im-plies that the case is more in, rather than out of, theset. For instance, if only well-informed stakehold-


ers accurately identify the state of practice inside afirm, this firm might be considered more in, ratherthan out of, the set of firms with potential to con-ceal their nonimplementation of policy. In contrast,if more—but not all—stakeholders accurately iden-tify a firm's state of practice, the firm might beconsidered more out of, rather than in, the set offirms with potential to conceal their nonimplemen-tation. Levels can be based on theoretical wisdomor in-depth knowledge of the cases (Rihoux &Ragin, 2008). We discuss below our calibration ofthe outcome and explanatory conditions.

Responses to Institutional Pressures

Drawing on social rating data, our executive in-terviews, and archival evidence to achieve triangu-lation of sources, we distinguished firms character-ized by a high and consistent degree of takingsubstantive action to implement CSR policy fromthose characterized by a low degree of substantiveaction.

We relied on social rating data for the initialselection of firms. Within each pair or triad, onefirm was more consistent in its implementationacross numerous dimensions than its peer(s).

All firms in our sample had adopted CSR-relatedpolicy. We attended to four instances of CSR policyadoption: the mention of societal benefit in missionstatements, the publication of ethical codes, report-ing social and environmental performance accord-ing to Global Reporting Initiative (GRI) guidelines,and membership in the UN Global Compact(www.unglobalcompact.org). Stakeholders identi-fied these practices as easily adopted without ne-cessitating implementation. As Weaver et al. (1999)suggested, codes and policy communications donot necessarily imply changes to internal behav-iors. Although CRI guidelines stipulate that firmsreport performance on 24 social and environmentalcriteria, firms monitor their own performance (GRI,2000). Likewise, though Global Compact signato-ries agree to implement a range of responsible prac-tices, "This does not mean that the Global Compactrecognizes or certifies that these companies havefulfilled the Compact's principles" (Clobal Com-pact, 2000). All firms in our sample mentionedsocial benefit in their mission statements, all pub-lished ethical codes, and all reported social andenvironmental performance following CRI guide-lines (at least, in part). Sixteen firms were membersof Global Compact; the one firm that was not amember (SMART) scored high on the implementa-tion of CSR policy.

In assessing substantive action, we attended tothe influence of CSR on firms' priorities, structure

of authority, decision making, and operations. Weselected these dimensions because they are compo-nents of the core of an organization (Hannan &Freeman, 1977). Specifically, we assessed the ex-tent to which firm-level performance targets in-cluded social responsibility-related criteria (andprogress toward these targets in the 2005-06); theinclusion of social responsibility-related criteria inthe performance appraisal of individual managers(targets and appraisal systems potentially reflectorganizational goals more accurately than publicmission statements); and the integration of CSR instrategic decision making and in operatingprocesses.

Table 2 outlines the influence of social responsi-bility along these dimensions in the 17 firms. Bettersocial performers show greater evidence of makingadaptations to their core than decouplers. Integra-tors have more stringent social responsibility-related firm-level targets than decouplers. The useof responsibility-related management appraisal cri-teria is more extensive in the integrators.

We allocated full membership (1) to firms thatwere consistent in their implementation and that,on average, had high scores on the dimensions weexamined. In particular, all of these firms had highscores for their integration of CSR in their opera-tions. We allocated nonmembership (0) to organi-zations that did not implement policy consistentlyand had, on average, low scores on the dimensions.However, decoupling is not necessarily binary (Yo-shikawa, Tsui-Auch, & McGuire, 2007), and fivefirms were partially consistent in their implemen-tation, typically integrating CSR unevenly in theiroperations. We allocated partial membership (0.5)to these firms.

Explanatory Conditions

Drawing on our theoretical development, weconsidered four conditions that might influencefirms' responses. Two conditions—informationasymmetry and stakeholder consensus—reflect theenvironment facing a firm and were measured atthe industry level. Two conditions—managerialconsensus and perceived interests in implementingpolicy—relate to attributes of the firm.

Potential information asynunetry. We measuredthe potential for firms to conceal nonimplementa-tion. Reporting differs across sectors as defined byindustrial activity (Clobal Reporting Initiative,2000), and we wished to understand which sectorsoffered firms the greatest latitude to concealnoncompliance.

Stakeholders identified corporations in the sec-tor they were familiar with and rated their social

2012 Crilly, Zoilo, and Bansen 1435

TABLE 2Evidence of Substantive Actions

CompanyCSR-Related

Performance TargetsCSR-Related Criteria inPerformance Appraisals

Integration of CSR intoStrategic Decision Making

Integration of CSR intoOperating Processes

SMART

TECHNIC

SILICON

NUCLEUS

FUSION

SCIENTIFIC

EXCAVATION

High: Environmental, socialtargets: progress tracked.

Medium: Enviromnental, socialtargets; few indicators.

Low: Few performance targets(only at country level).

High: Sustainability targets;progress tracked and audited.

High: Environment, health, andsafety targets.

Medium: Health, environment,and safety targets.

Medium: Some targets; progresstracked.

RESOURCE

MINER

POWER

ENERGETIC

Low: Some targets (judged laxby stakeholders).

Medium: Environment andsocial targets.

High: Performance goals(environment and social).

Hieh: Performance eoals(especially environment andsocial).

HIGHFINA;NÍGE Medium: Environmental andsocial targets; partialreporting.

GAPITAL Medium: Range of targets;progress not reported.

EDIBLE Medium: Wide range of goals;progress tracked.

CONSUMPTION Low: No environment, health,or safety targets.

DRUG Medium: Health, hygiene,environment targets; progressreported.

MEDICAL Medium: Targets for health andsafety; some tracking.

Medium: Applied accordingto domain.

Medium: Applied only tounits with specific CSRremit.

Medium: CSR criteria used toevaluate country managersonly.

Medium: Some link (extentvaries by function).

Medium: At least one CSR-related target for linemanagers.

Low: None at present.

Medium: Environment,health, and safety criteria(extent varies).

Low: Only enviromnentalcriteria for few employees.

Low: Only for few employees.

High: Most employees'salaries linked to socialperformance.

Medium: Some salarieslinked to socialperformance.

Medium: Some influence onperformance appraisals.

Low: None at present.

Low: No ties to CSR in mostmanagers' bonus.

Low: No link between socialperformance andcompensation.

High: Links to managers'bonuses.

Medium: No general link(exceptions for specificroles).

High: Highly integrated infunctions (e.g., innovation,procurement).

Medium: Influence on productmarket strategies(development, branding).

Low: Limited influence of GSRunit, as part of legal affairs,on strategy.

High: Strategies developed inline with social andenvironmental criteria.

Medium: Influence ofenvironmental and socialissues on major decisions.

Low: Limited integration.

High: Strong influence oncompetitive and corporatestrategy.

Medium: Ad hoc integration;some influence on productmarket strategies.

Low: Limited integration ofCSR in strategic decisions.

High: Focus on renewableenergies; early commitmentand follow-through.

Medium: Some influence onproduct market strategies.

High: Integration via localcommittees to tailor practiceto stakeholder needs

Medium: Limited role for CSRfunction; GSRrepresentatives in businessunits.

High: Integration into productdevelopment, sourcing,partnering.

Medium: Role when formingstrategy, but largely liniitedto branding.

High: Influence on R&D,investments, product andgeographic expansion.

Low: Low payback periodcriterion (two years) forinvestments acts as barrier.

High: Extensive integration infunctions (e.g., HR, supplierrelations).

Medium: Integration in somefunctions (e.g., HR,environment, procurement).

Low: Low integration butdeveloping as a result of theefforts by legal affairs.

High: Extensive integration inprocurement andmanufacturing.

Medium: Some, but uneven,integration in finance andpurchasing.

Medium: Narrow scope ofintegration (pollution andemissions monitoring).

High: Standardized social andenvironmental approachesacross firm.

Medium: Efforts (uneven acrossfirm) to integrateenvironment and safetypolicy.

Medium: Varyingimplementation across sites.

High: Sustainability-relatedresponsibilities in eachfunction.

Low: Gentralized GSR office;limited integration inbusiness functions.

High: Social impacts ofcustomers' activities takeninto account in lending.

High: Social impacts ofcustomers' activities takeninto account in lending.

High: Integration intoprocurement, HR, etc.; samestandards across all units.

Medium: Environment, healthand safety system in 50percent of units; few audits.

High: Extensive integration ofGSR into clinical trials,operations, marketing.

Medium: Inconsistent acrosslocations.

performance on a 1-10 scale. We obtained 420 eval-uations, 302 of which were of firms also rated byInnovest. We converted the Innovest social perfor-mance scores to the same 1-10 scale. We measuredthe discrepancy by subtracting the transformed In-

novest scores for firms in each sector from thestakeholder evaluations of the same firms. Large,positive gaps imply that stakeholders are more gen-erous than a rating agency with access to moreobjective data.

1436 Academy of Management fournal December

Firms in the banking and food sectors consis-tently received higher evaluations from stakehold-ers than their substantive performance, as mea-sured by social ratings, warranted. In explainingtheir evaluations, stakeholders drew attention tostandards and frameworks, such as Global Com-pact, whose signatories monitored their own com-pliance. Firms in these sectors can easily buildsmokescreens around their internal practices. Theirsustainability reports contain few quantitative data,impeding comparisons. Correspondingly, we arguethat these sectors offer high potential for informa-tion asymmetry, and we allocated full membership(1) to firms in these sectors.

At the other extreme, stakeholders had the mostinformation about firms in the chemicals and high-technology sectors. The data most relevant to socialperformance in the chemicals sector, which con-cern environmental emissions, are publicly acces-sible and enable comparisons among firms. In thehigh-technology sector, stakeholders are well in-formed, using technology to access data and toexchange information. Both sectors have agenciesto monitor social performance, restricting firms'latitude to engage in impression management. Cor-respondingly, the potential for information asym-metry is low, and we allocated nonmembership (0)to firms in these sectors.

In the remaining sectors, stakeholders used bothobjective and subjective criteria in evaluating or-ganizations. In the Pharmaceuticals sector, report-ing requirements are high, and the gap betweenstakeholder evaluations and the Innovest ratings issmall. We thus allocated a low degree of member-ship (0.33) to the pharmaceutical sector, refiectinga positive, but marginal, potential for informationasymmetry. In contrast, the larger gap betweenstakeholder evaluations and Innovest scores in thenatural resources and energy sectors impliesgreater information asymmetry. In both sectors, in-terviewees were largely unaware of practices out-side of their home countries. On average, stake-holders here did not distinguish between firmswith high and low social performance. We attrib-uted partial membership (0.66) to the firms in thesesectors, representing a comparatively high poten-tial to confuse stakeholders.

Stakeholder consensus. We assessed stake-holder consensus in expectations about firms' so-cial engagement across the sectors in our study.

We derived three disfinct perspectives on socialresponsibility from om- coding of stakeholders' an-swers to the interview question, "What is the corpo-rate responsibility of firms within sector X?" Stake-holders differed in the scope of engagement expectedfrom firms, from one premised on narrow engage-

ment focused on fulfilling fiduciary duties and legalobligafions to one premised on broad engagementfocused on the solution of societal problems such ashealth and poverty. Between these extremes, therewas support for moderate engagement to createwealth for, or minimize damage to, the constituenciesthat firms directly affect. We display an overview ofthese perspecfives in Table 3 and link them to per-spectives in the extant CSR literature.

We developed a detailed coding lexicon, an ex-tract of which is provided in Table 3. Two research-ers independently coded statements emphasizingfiduciary duties to shareholders and compliancewith legal requirements as legal-economic (e.g.,"The main responsibility is the duty to sharehold-ers"). Statements focused on responsibility towardsstakeholders affected by firms' activities werecoded as counterpart-centric (e.g., "Firms shouldpay attention to the well-being of customers, em-ployees, communities and others impacted by theiractivities"). Statements emphasizing broader re-sponsibilities, such as the solution of societal prob-lems around the environment, poverty, and educa-tion, were coded as citizenship-centric (e.g., "Firmshave a role to contribute to the betterment of soci-ety by driving environmental improvements").Where a stakeholder voiced multiple perspectives,we focused on the perspective that the stakeholderprioritized. We assessed interrater reliability usingKrippendorfPs alpha, which controls for chancecoincidences (Krippendorff, 2004). The alpha coef-ficient was 0.81, which is above Krippendorffsmost conservative threshold for reliability. We re-solved discrepancies through discussion.

We scored the extent of social engagement con-sidered appropriate for each industry by its stake-holders before measuring the variance across thestakeholders in each sector. Economic-legal state-ments, reflective of the narrowest form of corporateengagement, were scored 1. Counterpart-centricstatements were scored 2. Citizenship-centric state-ments reflective of the most extensive corporateengagement were scored 3. We then measured thevariance across stakeholders in each sector.

Variance was lowest in the extraction (0.40) andenergy (0.42) sectors, reflecting consensus aroimdbroad engagement, and we coded organizations inthese sectors as fully in the set of organizations facinghigh stakeholder consensus (1). According to our in-terviewees, negative externalities in resource-inten-sive sectors enable stakeholders to coordinate inifia-tives to set the corporate agenda, reinforcingcollective frames of acüon. Variance was highest inthe high technology (0.73), banking (0.73), and food(0.72) sectors, implying ambiguity surroimding socialresponsibilifies here. We coded firms in these sectors


TABLE 3Competing Perspectives on Corporate Social Engagement

Characteristic andExamples lensen and Meckling (1976) Freeman (1984) Margolis and Walsh (2003)

Logic of corporatesocial responsibility

Aim

Manager statements

Stakebolder statements

Statements coded(extracts)

Legal-economic: Responsibilitydefined in terms of legal andeconomic constraints onbebavior

Maximize sbarebolder value undersocial and environmentalconstraints

"Our responsibility is to bave anopen and transparent relationwitb tbe tax autborities."

"A company's responsibility is togenerate wealtb, and not topromote social welfare. Itsresponsibility is to survive."

"Responsibility is to maximizesbarebolder wealtb and live upto national regulation."

"It's about compliance, obeyingtbe laws and being transparent."

Optimizing tax paymentsLegal complianceCompetitiveness/growtbNot to promote social welfareMaximize/create sbarebolder valueMake moneyConduct business/deliver productsTransparency

Counterpart-centric: Responsibilitydefined in terms of obligationstowards stakeholders directlyaffected by tbe firm's activities

Expands to encompass network ofdifferent actors witb ties to thecorporation

"We sbould meet all stakebolderexpectations wbile satisfyingsbarebolder expectations at tbesame time."

"Companies are required to respondto tbe demands of internal andexternal stakeholders."

"Corporations bave responsibilities tocommunities where they operate,to their shareholders, to theiremployees and to the public interms of product liability."

"The firm is an organization withresponsibilities towards multiplestakeholders."

Meet all stakeholder expectationsDialogue with stakeholdersReduce environment impacts/manage

ecological riskMinimize impacts on immediate

stakeholdersProvide secxure employmentFair wages for employees

Citizenship-centric:Responsibility defined in termsof social impact, touching thosewho have no economic or legal

• rights over tbe firmExtends responsibility further,

potentially encroaching intospaces where goverrunents areactive

"Our responsibility is to use ourknowledge and skills for theimprovement of human life.We want to impact the world."

"Take a framework like theMillennium DevelopmentGoals. I see the contributionswe have to complete projectsserving society."

"Companies should try to leavethe world a better place—byrestoring the environment andproviding livelihoods."

"Firms have a politicalresponsibility. Social problemshave to be managed bycompanies, NGOs, andgovernments."

Improve life/well-being/socialvalue

Reduce povertyServe societyImprove environment (beyond

restricting damage by ownfirm)

Improve healthContribute to educationMeet needs of developing

countries

as fully out of the set of firms facing stakeholderconsensus (0). Stakeholders active in these sectorsoften identified posifive externalities—for example,facilitating access to education and health services.There was less consensus on firms' responsibility to"do good" than there was on their responsibility tominimize harm. Variance lay between these two ex-tremes in the remaining sectors, and we calibratedfirms in these sectors as facing neither high nor lowstakeholder consensus (0.5).

Organizational interests. Some firms expresscommitment to socially approved policies to securelegitimacy, which can involve impression manage-ment (Bansal & Clelland, 2004). Others perceiveengaging in socially approved behaviors as compet-itiveness-enhancing (Bansal & Roth, 2000). Thesemotivations could shape how executives represent

the returns from actions to satisfy stakeholders. Inthe interviews, executives explained their ratio-nales for social engagement and distributed tenpoints between four statements. Two statementsinvolved risk reduction and cost cutting ("It re-duces firm risks" and "It reduces costs"), whereastwo statements involved opportunity maximization("It helps to sell more" and "It is a source of newopportunities").

We calculated the ratio of opportunity maximi-zation to risk reduction in rationales by aggregatingthe responses of the executives in each firm. Exec-utives in five firms predominantly perceived CSRas a tool for generating business opportunities.These firms were coded as fully in (1) the set oforganizations with a focus on opportunity maximi-zation in driving CSR. Four firms clustered at the


other end of the spectrum, emphasizing risk reduc-tion approximately twice as prominently as oppor-tunity maximization, and were coded as fully out-side (0) the set. Of the remaining firms, two werejust short of prioritizing opportimity maximizationand were coded as predominantly in (0.66) the set.Six emphasized risk reduction and were coded aspredominantly outside (0.33) the set.

Managerial consensus. We measured the con-sensus of managers on how they represented thesocial responsibility of their firms. To do so, weused answers to the same question that we askedstakeholders (described above) and an identicalcoding procedure. Krippendorff's alpha was 0.80.Firms clustered into three groups. Seven firms hadvariance lower than 0.3 and were coded as mem-bers (1) of the high-consensus set. Executives atNUCLEUS, the firm just below this threshold of 0.3,expressed fairly consistent definitions of theirfirm's responsibility. Despite prioritizing differentfoci of CSR policy (e.g., health and safety, environ-ment, or education), together they emphasized theimportance of a shared understanding of the corpo-rate purpose (e.g., "Agreeing what is important can-not be done in a decentralized way. You do itcentrally close to the top. Once you have thatagreed, you can roll it into the organization."). Atthe other extreme, six firms had variance higherthan 0.6 and were coded as nonmembers (0). Exec-utives at DRUG, the firm just above this thresholdof 0.6, emphasized the lack of common imderstand-ing (e.g., "All our units are very decentralized. All

of a sudden we're realizing that stakeholder rela-tions need to be in better harmony because theworld doesn't see us as these separate functions").The remaining firms were coded as partial mem-bers (0.5). Neither consensus nor dissensionaroimd the fundamental responsibility of the firmemerged as an important theme in our interviewswith their executives.

In Table 4 we display the calibration of all 17 casesacross the outcome and explanatory conditions.

Analysis

The second step in performing an fsQCA involvesthe construction of a truth table to identify configu-rations of conditions associated with an outcome.The truth table lists all two-to-the-fourth (2*) logicallypossible configurations. We deleted configurationsnot associated with any of the 17 firms. We thenspecified a consistency threshold to select the config-urations reliably associated with one of the outcomes.Consistency, which can range from 0 to 1, measures"the degree to which instances of an outcome agree indisplaying the causal condition" (Ragin, 2008: 44).One guideline is to select a threshold that corre-sponds to a break observed in the disfribuüon ofconsistency scores (Schneider, Schulze-Bentrop, &Paimescu, 2010). Following this approach, we ap-plied a threshold of 0.748.

The next step involves an algorithm to simplifythe configurations and to arrive at a more parsimo-nious understanding. We employed the truth table

TABLE 4Calibration Table for Fuzzy Set Qualitative Analysis"

Firm

SMARTTECHNICSILICONNUCLEUSFUSIONSCIENTIFICEXCAVATIONRESOURCEMINERPOWERENERGETICHIGHFINANCECAPITALEDIBLECONSUMPTIONDRUGMEDICAL

Potentialfor

Asymmetry

0000000.660.660.660.660.6611110.330.33

StakeholderConsensus

0000.50.50.51111100000.50.5

OrganizationalInterest

110.660.330.330.330.330.33010.330.330.661010

ManagerialConsensus

10.50110110.50.5010.50001

SubstantiveAction

10.5010.5010010.510.51010.5

" We input values of 0.5 as 0.499 in the fs/QCA software program (Ragin, Drass, & Davey, 2006). This is necessary because cases withcondition values of 0.5 are automatically dropped during the analysis.


algorithm (Ragin, 2008), which uses counterfactualanalysis to speculate about the plausible outcomesof the logically possible combinations that do notexist in the data set.^ The most parsimonious solu-tion contains only core conditions that have thestrongest evidence linking them to the outcome.This parsimonious solution is contained within anintermediate solution. The intermediate solution ismore conservative because it only takes advantageof the most plausible simplifying assumptions(Ragin, 2008). In addition to core conditions, itcontains conditions that are considered peripheral.

RESULTS

In Table 5 we display the configurations of con-ditions associated with the decoupling of CSR pol-icy from practice as well as those that predict con-sistent implementation. Each column represents adistinct configuration. Our table follows the ap-proach of Ragin and Fiss (2008) by displaying theintermediate solutions consisting of core condi-tions and peripheral conditions. The peripheralconditions are represented by smaller symbols thanthe core conditions.

The table shows evidence of set-theoretic rela-tionships. The overall solution that explains decou-pling has a consistency of 0.87, and the solutionthat explains implementation has a consistency of1. Prior research considers consistency scores of at

^ For more information on counterfactual analysis,please consult Ragin (2008) or Fiss (2011).

least 0.8 acceptable (Fiss, 2011). Coverage, whichranges from 0 to 1 (Ragin, 2008), measures theextent to which the solutions explain all cases ofdecoupling and implementation. The coverage ofthe solution explaining decoupling (0.58) is lowerthan that of the solution explaining implementa-tion (0.67), implying that our implementation solu-tion is empirically more powerful. We also reportmeasures of consistency and coverage for each in-dividual configuration.

We found two configurations associated with de-coupling and two configurations associated withimplementation. To shed light on the nature ofresponses, we identified the firms that most consis-tently exhibit the characteristics of each configura-tion and provide evidence in Table 6 from ourinterviews with executives and stakeholders. Weselected firms with a membership of at least 0.5 inthe respective configuration. Membership in a con-figuration is equal to the minimum degree of mem-bership in any condition that contributes to theconfiguration. Cenerally, no case can have a mem-bership score greater than 0.5 in more than oneconfiguration (Ragin, 2008).

We discuss first the configurations associatedwith decoupling before turning our attention toimplementation.

Explaining Decoupling

Our analysis reveals two pathways to decou-pling. Each configuration's total and unique cover-age scores are identical (Table 5), implying theuniqueness of each pathway.

TABLE 5Configurations of Causal Conditions Leading to Decoupling and Implementation"

Configurations forDecoupling Configurations for Implementation

Causal Conditions

External environmentPotential for asymmetryStakeholder consensus

FirmOrganizational interestManagerial consensus

ConsistencyRaw coverageUnique coverageOverall solution consistencyOverall solution coverage

1

?

OO

0.940.310.31

2

Oo

O

0.800.270.27

0.870.58

3

f

fO

1.000.280.24

4a

O

o

f

1.000.320.161.000.67

4b

O

of

1.000.260.07

" Key: í = core causal condition (present). í = peripheral causal condition (present). O = core causal condition (absent), o =peripheral causal condition (absent). This format of presenting the results from the QCA is based on Ragin and Fiss (2008).

1440 Academy of Management Joumal December

TABLE 6Qualitative Evidence"

Organization Managerial Evidence Stakeholder Evidence

Configuration 1 ; Evasive decouplingCONSUMPTION There's an opportunistic way of looking at it. We have

a good communications department. We havereceived reputation awards but need to change ourinternal behavior.

MINER We give little weight to health, safety and environmentconsiderations. The pendulum is shifting but it'sslow because we are not encouraged by the marketsto be proactive in our change process.

ENERCETIC It's exploitation in the line of "I do things well and Itell them even better."

Configuration 2: Emergent decouplingSILICON There are different motivations and understandings

from different employees. . . . You have to keeptesting internally and externally to know whenyou're doing enough or too much.

TECHNIC You will never hear anyone within TECHNIC say thatwe're there. Society changes, norms change.Companies have to constantly evolve and see whatthe demands of society are.

SCIENTIFIC The big issue is consensus: people being clear aboutwhat they want. Should our activity be society- orbusiness-focused? It is a conceptual challenge. Thesupply-side stuff becomes easier if you are clear whatthe demand side is.

Configuration 3: Strategic implementationEDIBLE You need to build into your processes the certainty that

you can carry on in the long term—developing thenext generation, securing materials, or managing yoiufootprint. Those things are better managed when it'sin the company's self-interest [to do so].

POWER In addition to complying with the law, you have to takecare of the interests of other groups such as clients,shareholders, the environment, employees, supphers,public administration, society.

Configuration 4: Routine implementationSMART It was ingrained in our culture from the day I joined. It

is not something we questioned. There have alwaysbeen programs here for providing assistance to thecommunity, for providing talent and people to workin the community.

HIGHFINANCE CSR is a crucial element to create homogeneity, to keepthe group together in spite of the many differenceswithin it. So there is an identity-based motivation forengaging in CSR.

NUCLEUS Agreeing what is important cannot be done in adecentralized way. You do it centrally close to thetop. Once you have that agreed, you can roll it intothe organization.

It would be wise if they could separate themselvesfrom the pack, not just for the purposes of publicrelations, but for real.

They are doing a lot of horrible things and a fewsomewhat good things. Yet they're making a greatfuss about these small things that they do good. It'sjust a lot of blah-blah.

For them, CSR has been a communication choice. Therehasn't been engagement with these issues within thefirm, but ENERGETICS communication is veryeffective.

It really depends on the individual you're dealingwith. SILICON is limited by their own managers. Iencounter variance in how richly they xmderstandwhat their responsibilities are.

We don't know if they're applying their businessprinciples overseas. I visited a couple of theirfactories in China. Even they don't have enoughinformation about the circumstances there.

SCIENTIFIC does not walk the talk. Decision makershave no consequences for their behavior, and thefeedback mechanisms are poor due to the highmobility of individuals. There is no chain of

responsibility.

They really think about what they are doing. It is notan afterthought. They have a close relationship toclients. This is what CSR is about.

We interact with many firms in the energy sector.POWER is the one that has tried most to deepen therelationship. It also provides us with the mostinformation. We are further in our relationship withthem than with other energy providers.

In SMART, like most companies who really do CSR, thegeneral manager believes in this. It cannot be a task ofthe director of communication. You need power toinfluence. When you join, you have to sign theconduct guidelines and behave accordingly.

The firm is highly committed to CSR as a way to dobusiness. It has a long history on CSR. There'sdirection from the top, and there is also a sharedcode of conduct.

They lack a strategic mandate. [But] looking at theirprocesses, they're good. The way they deal withsupply chains is good. They have a proactiveculture in addressing relevant issues.

' Cases are assigned to configurations on the basis of their membership of at least 0.5 in the configuration.

Evasive decoupling. Configuration 1 impliesthat, when firms can conceal a lack of implemen-tation, and when their managers frame the rationalefor CSR in terms of risk reduction, they do notimplement policy consistently. Further, these cases

of decoupling occur in conditions of low consensusamong managers, but not necessarily low consen-sus among stakeholders.

Three cases belong to this configuration: CON-SUMPTION, ENERCETIC, and MINER. Our inter-


view evidence overwhelmingly points to their non-implementation of CSR policy as a deliberatechoice. Stakeholders accused these firms of "play-ing the CSR game" and intentionally trying to con-ceal nonimplementation.

It's a part of [GONSUMPTION's] marketing strategyto have a good social image as a company. [The aimis] to have liberal laws on sale, on trade, on market-ing and to keep those liberal laws.

One of the key criticisms against MINER is tbat theyuse the typical "divide and rule" tactic of workingwitb certain people within a community, cleverlygenerating their support. They find support fromone part of the community rather than the otberpart. . . . They use initial agreement to wear downresistance from others.

This intentional adoption of policy to convinceexternal constituencies is consistent with the out-ward-facing roles that the CSR units occupy in thethree firms. In ENERGETIC, the CSR tmit forms partof the communications department. In CONSUMP-TION and MINER, the CSR units fall under theremit of corporate relations.

We note that managerial consensus is low eventhough different stakeholders share similar expec-tations about corporations. Individual managerswho wish to drive implementation are constrainedby the lack of resources allocated to CSR imple-mentation because it is perceived as an tinneces-sary cost. The executive in charge of compliance atMINER explains the restrictions he faces becausethe management team as a whole views investing inCSR as deleterious to financial performance: "Ihave to encourage colleagues to be more responsiveto safety and the environment. I would love to seethat, but because the bottom line is strictly finan-cial, those concerns are out of scope."

Emergent decoupling. Configin-ation 2 predictsdecoupling when managers disagree in their under-standing of CSR and face stakeholders with con-flicting expectations. Because this decoupling canoccur even when firms are unable to conceal theirnonimplementation, it is at odds with the feignedcompliance traditionally associated with decou-pling as a strategic choice. Low stakeholder consen-sus—for example, in the high-technology and foodsectors—creates uncertainty for managers in eval-uating alternative responses. However, this uncer-tainty is insufficient to explain decoupling. Rather,these cases of decoupling also involve low consen-sus among managers. Conflicting stakeholder de-mands become reflected in dissension within man-agement teams, and this internal dissensionfacilitates resistance to implementation and asearch for local solutions.

Three cases, SILICON, TECHNIC, and SCIEN-TIFIC, belong to this second configuration. We notethat these firms have arranged their CSR functionsin different units. In SILICON, CSR is the remit ofeach regional unit. In TECHNIC, product units,businesses, and regions have their own remit. InSCIENTIFIC, individual business units have sub-stantial autonomy, and the primary function of theCSR department is to report on social and environ-mental performance. A consequence is that, despitecorporation-wide policy in each firm, implementa-tion is often left to the discretion of business unitand line managers.

Despite these structural arrangements, we findlittle evidence of decoupling as an intentional re-sponse in these firms. As Table 6 suggests, thecorporate headquarters do not purposely set out toavoid implementation or to deceive stakeholders.Rather than the result of intent, the decoupling inthese cases is the result of the emergence of localbehaviors that are sometimes at odds with officialpolicy. A SCIENTIFIC stakeholder contrasted the"nice, long-haired development people" with"those people on the ground." Similarly, TECHNICstakeholders are often better informed than head-quarter managers about the state of implementationoutside of Etirope. One representative of an envi-ronmental group, who had visited TECHNICS fac-tories in China, remarked, "Even they (the head-quarter staff) don't have enough information aboutthe manufacturing circumstances there."

Although all three firms fit the same configura-tion, we note one distinction. In SILICION, whereregional executives oversaw implementation, in-consistencies in implementation appeared toemerge from local discretion to interpret policy.Inconsistent stakeholder expectations in the infor-mation technology sector allowed executives tohighlight contextual pressures. For example, a re-gional head of SILICON, who emphasized "differ-ent understandings and different expectations onconipanies, depending on the market in which theyoperate," was able to withdraw a project to collab-orate with a ministry of education, justifying thedecision by a lack of local acceptance. In contrast,in SCIENTIFIC and TECHNIC, inconsistencies inimplementation emerge from the imperfect replica-tion of CSR practices across fragmented environ-ments. The executive in charge of CSR at SCIEN-TIFIC emphasized the intent to secure uniformcompliance: "I can't have two sets of standards in amultinational company, and therefore my task is toprovide increased awareness of standards acrossthe business models." Yet implementing policyconsistently becomes challenging as corporationsexpand and face new stakeholders. A TECHNIC


stakeholder described key challenges facing middlemanagers, who were often excluded from policydiscussions and training:

TECHNIC has awareness of CSR issues, but mana-gerial skills are more relevant—not for top manage-ment because tbey establisb policies, but for middlemanagers who implement them. Tbe ones in chargeof daily operations need to be made aware and needmore meuiagerial skills to understand how to mea-sure performance on tbese issues.

Explaining Implementation

We performed a second analysis to identify theconditions associated with the consistent imple-mentation of policy. Our analysis, shown in Table5, reveals two pathways.

Strategic implementation. Configuration 3 high-lights organizational interests related to CSR. Man-agers implement CSR policy consistently whenthey link CSR to potential business opportunitiesand when information asymmetry would otherwiseimpede the establishment of close stakeholder re-lationships. EDIBLE and POWER belong to thisconfiguration. Both firms emphasize CSR as asource of potential opportunity, premised on secur-ing access to resources from stakeholders, includ-ing know-how from suppliers and licenses to oper-ate from communities and governments. As anEDIBLE executive stated, "There are compellingreasons to engage in CSR, like attracting and retain-ing the kind of people that you want to be workingfor you and with you, and the ability of business toshare expertise with the NGO sector." Though thepotential for asymmetry sometimes prompts gam-ing, as firms can conceal their lack of implementa-tion, asymmetry also raises distrust among stake-holders. Under these conditions, stakeholders arereluctant to engage with corporations because theyare less able to tell the good from the bad and mightlend legitimacy to corporations with poorimplementation.

The evidence in Table 6 underscores the trust thatEDIBLE and POWER had gained from their externalstakeholders. This trust appears to be crucial for themaintenance of productive firm-stakeholder relation-ships. In the food and energy sectors, organizationscan easily conceal their true states of practice. Stake-holders are thus wary of lending legitimacy to corpo-rations that, they suspect, have adopted CSR policy asa marketing device. The representative of a commu-nity group commented on the difficulty of evaluatingthe state of practice in CONSUMPTION, EDIBLE'Scompetitor:

We are not auditors. I talk to companies about thepolicies they have at tbe beadquarter level, but I

don't dare to think about the implementation. . . .When dialoging with some competnies, I have to becareful. It is always sort of a strategic game.

Routine implementation. Our second pathwayto implementation (configurations 4a and 4b) doesnot depend on a strategic interest in implementingpolicy. Instead, it rests on managerial consensusabout corporations' social responsibility. Internalconsensus matters when stakeholders disagreeabout the appropriate extent of corporate socialengagement and, hence, may be less able to coor-dinate action against firms' noncompliance. Thisimplies that the impetus for policy implementationin these cases comes from vríthin a firm rather thanfrom external pressure and that organizational val-ues or identity resolve uncertainty about the envi-ronment. Though configuration 4a includes the ab-sence of information asymmetry as a peripheralcondition, implying it is difficult for firms to con-ceal noncompliance, configuration 4b includes theabsence of organizational interest linked to imple-menting CSR policy as a peripheral condition. Pe-ripheral conditions show weaker evidence for acausal relationship with the outcome (implementa-tion), but the absence of organizational interest isparticularly revealing as it contrasts with its pres-ence in configuration 3.

SMART, HICHFINANCE, and NUCLEUS belongto this configuration. SMART executives referredconsistently to an internal impetus for implemen-tation. Managers described implementation asdriven by internal rules of behavior that appearedappropriate irrespective of their direct conse-quences for financial performance. A manager ofSMART, similarly facing weak stakeholder expec-tations for corporate social engagement, describedhow implementation was not merely the result ofcompliance with external pressures:

It [CSR] was ingrained in our culture from tbe be-ginning. [Our founder) tried to ingrain social valuesand to demonstrate to employees how important itwas not just to have a job but to try to influence tbesocial agenda. . . . From the first day I was withSMART we were instructed in this, and tbis is notsometbing we questioned.

Similarly, a NUCLEUS executive emphasizedthat institutional pressures led to the articulation ofCSR policy that was based on practices that hadalready been commonplace in the firm for manyyears:

We came up through a company witb a strong rep-utation in many of tbese areas. As long as I remem-ber, we were known as a company where peopleinternally were treated well, and they behaved de-cently in society. This has been a value for many


years. At some stage it was decided to formalize thisinto a triple bottom line because we were forced toreport.

All three firms internally resolved the uncertaintyinherent in their external environments. The unitsresponsible for CSR had an integrative function,formalizing policy across each of these firms.HIGHFINANCE had a single unit responsible forCSR that maintained links to all core business func-tions. SMART had a single council combining rep-resentatives from the CSR unit and its core businessfunctions. CSR programs were budgeted centrally.NUCLEUS had a cross-functional group for CSRincluding executives from different functions andgeographies that set "global guidelines and globalstandards and minimum requirements that countfor all."

In explaining implementation, we note the dis-crepancy between the configurations' unique cov-erage (0.28, 0.32, 0.26) and their total coverage(0.24, 0.16, 0.07). This discrepancy implies that ourexplanations for implementation are not mutuallyexclusive. The same firm could implement policyconsistently both because its managers identified astrategic rationale for doing so and because stronginternal norms favored implementation.

Finally, the firm CAPITAL is an interestinganomaly. At the time of our interviews, CAPITALhad a mediocre record of implementation, as evi-denced by poor social performance ratings, but itpossessed the characteristics of the strategic imple-mentation configuration (Table 4). We note thatCAPITAL did not formally adopt CSR policy until2004, substantially later than the other firms in oursample. Though implementation was, according toa stakeholder, "not yet completely integrated in theorganization and its operations," later social perfor-mance data suggest increased compliance. This im-plies that even intentional moves to articulate andimplement policy can take time to bear fruit, withimplementation potentially requiring ongoing ef-forts to change organizational processes.

DISCUSSION

The purpose of this study was to understand whyfirms in the same environment respond differentlyto institutional pressures for CSR. Not all cases ofdecoupling and implementation are identical.Firms decouple their behavior from stated commit-ments not only for intentional, exploitative rea-sons, but also as a result of uncoordinated, explor-atory attempts to respond to diverse and conflictingdemands in a generally well intended "muddlingthrough" process. Firms implement policy either

when executives collectively perceive instrumentalbenefits from doing so or when they resolve theuncertainty in their external environment by devel-oping an organizational view of the global set ofissues that they face and coordinating policyimplementation.

Multilevel Explanation of Institutional Response

Our first contribution to institutional theory is amultilevel explanation of response to stakeholderpressures that challenges and refines existing the-oretical accounts. Factors external and internal to afirm combine to influence when different re-sponses occur and when they are emergent ratherthan strategic attempts to exploit stakeholder igno-rance or to win stakeholder favor.

Most explanations of decoupling prioritize eithera firm's external environment or the internal organ-ization. Many investigations have addressed factorsthat are exogenous to firms, such as divergentstakeholder expectations (Meyer & Rowan, 1977)and the degree to which sectors are dominated bybeliefs and values rather than market forces (Scott& Meyer, 1983). Other investigations have ad-dressed the interests of executives inside firms(e.g., Westphal & Zajac, 1994). However, our resultssuggest that it is problematic to view forces at anysingle level of analysis as sufficient for decouplingto occur. For instance, the inability of stakeholdersto identify the state of practice in firms only pre-dicts decoupling if executives have unambiguousinterests against implementation. This is evidencedby CONSUMPTION and EDIBLE, two firms thatresponded differently to similar external stake-holder pressures.

The interplay between external environment andinternal organization matters for whether firms im-plement or decouple policy as well as how they goabout doing so. Decoupling may be more subtlethan a choice between implementation and evasion(Fiss & Zajac, 2006). Though existing theories ofdecoupling are largely silent on organizationallearning, under conditions of inconsistent stake-holder expectations, some firms engage in prob-lemistic search (Cyert & March, 1963), and decou-pling appears to emerge through variation in theirpractices. Emergence does not imply that thesepractices are entirely unintended (i.e., local man-agers might deliberately choose these practices),but rather that they are unintended by leadershipand at odds with official policy (Mintzberg & Wa-ters, 1985). Inconsistencies in implementation canfollow from different understandings of institu-tional pressures and also from the difficulty of rep-licating practices throughout a firm (Winter & Szu-


lanski, 2001). From this perspective, decouplingmight allow firms not only to reconcile competingforces in their environment, but also to accommo-date internal factions (Li & Hambrick, 2005; March,1962). The emergence of distinct practices as atruce among members with incongruent percep-tions and beliefs fits the evidence observed in firmssuch as SILICON, TECHNIC, and SCIENTIFIC.

We also find an explanation for implementationin similarly fragmented environments. Consistentimplementation of CSR policy is associated withmanagerial consensus about the social responsibil-ities of business. Conceivably, decoupling does nothappen when the dynamics of identity overtake thedynamics of problemistic search and imperfectlearning, because shared beliefs resolve uncertaintyand facilitate coordination and the replication ofpractices (Hardin & Higgins, 1996). Shared beliefscan sustain behaviors that are not motivated byapparent self-interest (Labianca, Gray, & Brass,2000), leading to convergence in behaviors withina firm.

Our identification of an emergent form of decou-pling and of an internally driven form of imple-mentation does not supersede strategic explana-tions. Rather, it specifies information asymmetry asa condition under which responses are likely to bestrategically motivated. Further, our behavioral in-terpretation contrasts with the depiction of decou-pling in Meyer and Rowan's (1977) seminal articlebecause our finding does not depend on complicitstakeholders who turn a blind eye to firms' actualpractices. However, the coverage score of our solu-tion (0.58) implies that our framework does notexplain all instances of decoupling. Other formsand drivers of decoupling may exist.

Microfoundations of Institutional Response

Our second contribution is to the emerging liter-ature on the microfoundations of institutional re-sponse (George et al., 2006; Tilcsik, 2010) and, inparticular, on the interaction of institutional pres-sures and cognition (Kennedy & Fiss, 2009). Atten-tion to individual executives and stakeholderscomplements research that recognizes a sociopolit-ical dimension to decoupling centered on powerand objective interests (Westphal & Zajac, 1998).

Identifying the microfoundations of response toinstitutional pressures is crucial to explain firmheterogeneity. We looked at two attributes of firms:how managers perceive their interests around CSR,and whether they agree in their understandings ofCSR. The firms in our study are complex corpora-tions, and their responses to institutional pressuresare not mechanistic (Delmas & Toffel, 2008; Ham-

brick, Finkelstein, Cho, & Jackson, 2005); however,ultimately individual actors exercise firms' discre-tion (Crilly, Schneider, & Zoilo, 2008). As institu-fional theory and the behavioral theory of the firmshare some similar assumpfions about actors'boiuided rafionality (Argote & Grève, 2007), opportu-nifies remain to link these research streams to iden-tify the microfoimdafions of insfitutional response.

An important link between institutional theoryand the behavioral theorj' of the firm concerns therole of coalitions (March, 1962) within and be-tween organizations in shaping responses. Prior re-search assesses the influence of internal actors onresponses (Westphal & Zajac, 1994) and the role ofexecutives in framing the environment for theirsubordinates (cf. Kaplan, 2008b). Our routine im-plementation configuration, exemplified by firmssuch as HIGHFINANCE and NUCLEUS, under-scores the importance of internal consensus for re-sponses. Our emergent-decoupling configuration,exemplified by SILICON, TECHNIC, and SCIEN-TIFIC, shifts attention to the role of stakeholders indriving implementation by communicating unam-biguous expectations. Conceivably, a lack of stake-holder consensus impedes coordination betweenstakeholders and risks, sending confusing signalsto managers.

We would welcome studies that investigate thecausal dynamics underlying the patterns we ob-served. Though our method, fsQCA, can identify"causal recipes" (Ragin, 2008: 23) for conditionsassociated with decoupling and implementation, itdoes not easily accommodate longitudinal data(Lacey & Fiss, 2009). Longitudinal and experimen-tal data would be necessary to ascertain whetherthe patterns we identified are truly causal. In par-ticular, longitudinal studies could illuminate theinteractions between strong stakeholder consensus,pressure on firms, and firm-level consensus or dis-sension over time. For example, though managerialconsensus predicts implementation in the absenceof stakeholder consensus, the relationship betweenmanagerial consensus and implementation is lessconsistent when stakeholders are united as to whatthey expect from firms. Our data firom the firmRESOURCE reveal that its executives have consen-sus around a narrow view of the firm's social re-sponsibility and allocate only limited resources to-ward implementing CSR policy. Though ourexplanation is tentative, we note that RESOURCEhas been a prominent target of activist campaigngroups. The unanimity within the managementteam is consistent with the cohesiveness thatemerges when individuals face a common enemy(Allport, 1958) and the activation of in-group iden-tities in response to external threats (Sherif, 1966).


Implications for Research into CSR

Our study also has implications for research intoCSR, much of which investigates executives' inten-tions, linking these to support for practices andpolicies. Yet intentions can be divorced from local-ized practice (Orlikowski, 2000), especially whenthe means for implementing policy are unclear(Coodrick & Salancik, 1996).

Two of our configurations suggest that some in-stances of decoupling and implementation resultfrom ongoing processes in a firm rather than stra-tegic decisions. Even if these responses are notnecessarily intended, this finding does not meanthat there is no role for active management or or-ganization design. Our finding of common struc-tures across firms fitting the same configiirationunderlines the importance of formal control to in-fluence resource allocation across dispersed organ-izations (Camoran & Dreèhen, 1986). Informal or-ganization also plays a role. The cognitions andmotivations of managers represent a barrier to im-plementation (Labianca et al., 2000). A tentativeconclusion from our study is that efforts to buildintemal norms and identity can facilitate the repli-cation cf practices in diverse units. This is consis-tent with evidence that shared heliefs ahout thecore purpose of a firm enahle coordination amongmanagers representing different functions and ge-ographi3s, allowing firms to attend simultaneouslyto competing stakeholders (Crilly & Sloan, 2012).

Our study has implications for our understand-ing of managers' rationales for CSR. Though ourtwo pathways to decoupling—the strategic and theemergent—are mutually exclusive, our two path-ways to implementation are not. By implication,implementation does not have to he only ahoutperceived strategic interest or ahout organizationalidentity and values. Rather, in some cases, theseseparate drivers of implementation appear to oper-ate simultaneously. The CSR literature distin-guishes instrumental motives from moral and rela-tional motives (Aguilera et al., 2007) and frequentlyviews noninstrumental motives as ptu:er (Quinn &Jones, 1995). Our finding provides tentative evi-dence that self-interest can coexist alongside othermotives. As firms implement CSR policy, their ex-ecutives may hecom^e sensitized to related husinessopportunities.

Finally, we note that implementation is notsolely a concern for executives. As evidenced hyfirms such as SCIENTIFIC, which lack an organiza-tional view of the CSR issues that they face, lowlevels of stakeholder coordination conceivablyreinforce executives' uncertainty about the appro-priate policies to adopt and implement. This un-

certainty is associated with pockets of local impro-visation as lower-level managers enjoy greaterdiscretion over implementation. Though imple-mentation is less consistent in these firms than intheir peers that forge a single organizational view,their local improvisation may be welcome to spe-cific stakeholders. Conversely, though stakeholdercoordination can constrain firms' attempts to de-couple, our findings also urge caution. If aggressivecampaigning causes executives to attend to riskreduction as a guideline for behavior, they maycomply with the letter of the law and allocate re-sources to political activities rather than undertakesuhstantive change (David et al., 2007). The recog-nition that executives sometimes muddle through,rather than set out with the intent to exploit theirstakeholders, is a useful starting point for firm-stakeholder dialogue.

REFERENCES

Aguilera, R. V., Rupp, D. E., Williams, C. A., & Ganapathi,J. 2007. Putting tbe S back in corporate social respon-sibility: A multilevel theory of socictl change in or-ganizations. Academy of Management Review, 32:836-863.

Akerlof, G. 1970. Tbe market for "lemons": Quality un-certainty and Üie market mechanism. QuarterlyJoumal of Economics, 84: 488-500.

Allport, G. W. 1958. The nature of prejudice. GardenCity, NY: Doubleday.

Argote, L., & Grève, H. R. 2007. A behavioral theory of thefirm—40 years and counting: Introduction and im-pact. Organization Science, 18: 337-349.

Axelrod, R. 1984. The evolution of cooperation. NewYork: Basic Books.

Bansal, P., & Glelland, I. 2004. Talking trash: Legitimacy,impression management, and unsystematic risk inthe context of the natural environment. Academy ofManagement Joumal, 47: 93-103.

Bansal, P., & Roth, K. 2000. Why companies go green: Amodel of ecological responsiveness. Academy ofManagement Joumal, 43: 717-736.

Basu, K., & Palazzo, G. 2008. Corporate social responsi-bility: A process model of sensemaking. Academy ofManagement Review, 33: 122-136.

Battilana, J., & Dorado, S. 2010. Building sustainablehybrid organizations: The case of commercial micro-finance organizations. Academy of ManagementJoumal, 53: 1419-1440.

Christian Aid. 2004. Rehind the mask: The real face ofcorporate social responsibility. London: ChristianAid.

Christmann, P., & Taylor, G. 2001. Globalization and tbe


environment: Determinants of firm self-regulation inGhina. Journal of International Business Studies,32: 439-458.

Gress, D. M., & Snow, D. A. 1996. Mobilization at themargins: Resources, benefactors, and tbe viability ofhomeless social movement organizations. AmericanSociological Review, 61: 1089-1109.

Grilly, D., Schneider, S. G., & Zoilo, M. 2008. The psy-chological antecedents to sociedly responsible be-havior. European Management Review, 5: 175-190.

Grilly, D., & Sloan, P. 2012. Gorporate attention to stake-holders: Enterprise logic and an inside-out explana-tion. Strategic Management Journal, 33: 1174-1193.

Gyert, R. M., & March, J. G. 1963. A behavioral theory ofthe firm. Englewood Gliffs, NJ: Prentice Hall.

David, P., Bloom, M., & Hillman, A. J. 2007. Investoractivism, managerial responsiveness, and corporatesocial performance. Strategic Management Jour-nal, 28: 91-100.

Davis, K. 1973. The case for and against business as-sumption of social responsibilities. Academy ofManagement Journal, 16: 312-323.

Delmas, M., & Toffel, M. 2008. Organizational responsesto environmental demands: Opening the black box.Strategic Management Journal, 29: 1027-1055.

Den Hond, F., & de Bakker, F. G. A. 2007. Ideologicallymotivated activism: How activist groups influencecorporate social chcinge activities. Academy of Man-agement Review, 32: 901-924.

Fedderson, T., & Gilligan, T. 2001. Saints and markets:Activists and the supply of credence goods. Journalof Economics and Management Strategy, 10: 149-171.

Fiss, P. G. 2011. Building better causal theories: Afuzzy set approach to typologies in organizationresearch. Academy of Management Journal, 54:393-420.

Fiss, P. G., & Zajac, E. J. 2004. The diffusion of ideas overcontested terrain: The (non)adoption of a share-holder value orientation among German firms. Ad-ministrative Science Quarterly, 49: 501-534.

Fiss, P. S., & Zajac, E. J. 2006. The symbolic managementof strategic change: Sensegiving via framing and de-coupling. Academy of Management Journal, 49:1173-1193.

Freeman, R. E. 1984. Strategic management: A stake-holder approach. Boston: Pitman.

Gamoran, A., & Dreeben, R. 1986. Goupling and controlin educational organizations: An explication and il-lustrative comparative test. Administrative ScienceQuarterly, 31: 612-632.

Garud, R., Jain, S., & Kumaraswamy, A. 2002. Institu-tional entrepreneurship in the sponsorship of com-mon technological standards: The case of Sun Mi-crosystems and Java. Academy of ManagementJournal, 45: 196-214.

George, E., Ghattopadhyay, P., Sitkin, S., & Barden, J.2006. Gognitive underpinnings of institutional per-sistence and change: A framing perspective. Acad-emy of Management Review, 31: 347-365.

Global Reporting Initiative. 2000. A common frameworkfor sustainability reporting, www.globálreporting.org.

Goodrick, E., & Salancik, G. R. 1996. Organizational dis-cretion in responding to institutional practices: Hos-pitals and Gesarean births. Administrative ScienceQuarterly, 41: 1-28.

Greening, D. W., & Gray, B. 1994. Testing a model oforganizational response to social and political is-sues. Academy of Management Journal, 37: 467-498.

Hambrick, D. G., Finkelstein, S., Gho, T. S., & Jackson,E. M. 2005. Isomorphism in reverse: Institutionaltheory as an explanation for recent increases in in-traindustry heterogeneity and managerial discretion.In B. Staw & R. M. Kramer (Eds), Research in organ-izational behavior, vol. 26: 307-350. Greenwich,GT: JAI.

Hannan, M. T., & Freeman, J. H. 1977. The populationecology of organizations. American Journal of So-ciology, 82: 929-964.

Hardin, G. D., & Higgins, E. T. 1996. Shared reality: Howsocial verification makes the subjective objective. InE. T. Higgins & R. M. Sorrentino (Eds.), Handbook ofmotivation and cognition: The interpersonal con-text, vol. 3: 28-84). New York: Guilford.

Jensen, M. G., & Meckling, W. H. 1976. Theory of thefirm: Managerial behavior, agency costs and owner-ship structure. Journal of Einandal Economics, 3:305-360.

Jones, T. M. 1995. Instrumental stakeholder theory: Asynthesis of ethics and economics. Academy ofManagement Review, 20: 404-437.

Kaplan, S. 2008a. Gognition, capabilities, and incentives:Assessing firm response to the fiber-optic revolution.Academy of Management Journal, 51: 672-695.

Kaplan, S. 2008b. Framing contests: Strategy making un-der uncertainty. Organization Science, 19: 729-752.

Kennedy, M. T., & Fiss, P. G. 2009. Institutionalization,framing, and diffusion: The logic of TQM adoptionand implementation decisions among U.S. hospitals.Academy of Management Journal, 52: 897-918.

King, A. A., Lenox, M. J., & Terlaak, A. 2005. The strate-gic use of decentralized institutions: Exploring cer-

2012 Crílly, Zoilo, and Hansen 1447

tification witb the ISO 14001 management stan-dards. Academy of Management Journal, 48:1091-1106.

Krippendorff, K. 2004. Content analysis: An introduc-tion to its methodology. Thousand Oaks, CA: Sage.

Kulkarni, S. P. 2000. Information asymmetry among or-ganizational stakeholders. Journal of Business Eth-ics, 27: 215-228.

Labianca, C, Cray, B., & Brass, D. J. 2000. A groundedmodel of organizational change during empower-ment. Organization Science, 11: 235-257.

Lacey, R., & Fiss, P. C. 2009. Comparative organizationalanalysis across multiple levels: A set-theoretic ap-proach. In B. King, T. Felin, & D. Whetten (Eds.),Research in the sociology of organizations, vol. 26(Studying differences between organizations: Com-parative approaches to organizational research): 91 -116. Bingley, U.K.: Emerald.

Leifer, E. M.. 1988. Interaction preludes to role setting:Exploratory local action. American Sociological Re-view, 53: 865-878.

Leland, H. E. 1979. Quacks, lemons, and licensing: Atheory of minimum quality standards. Journal ofPolitical Economy, 87: 1328-1346.

Li, J. T., & Hambrick, D. C. 2005. Factional groups: A newvantage on demograpbic faultiness, conflict and dis-integration in work teams. Academy of Manage-ment Journal, 48: 794-813.

March, J. C. 1962. The business firm as a political coali-tion. Journal of Politics, 24: 662-678.

Margolis, J. D., & Walsb, J. P. 2003. Misery loves compa-nies: Rethinking social initiatives by business. Ad-nmiistrative Science Quarterly, 48: 268-305.

Meyer, J. W., & Rowan, B. 1977. Institutionalized organ-izations: Formal structure as myth and ceremony.American Journal of Sociology, 83: 340-363.

Mintzberg, H., & Waters, J. A. 1985. Of strategies, delib-erate and emergent. Strategic Management Journal,6: 257-272.

Naj^yar, P. R. 1990. Information asjrmmetries: A source ofcompetitive advantage for diversified service firms.Strategic Management Journal, 11: 513-519.

Novo Nordisk. 2011. Novo Nordisk annual report 2010.http://annualreport.novonordisk.com/. AccessedJanuary 1.

Orlikowski, W. J. 2000. Using technology and constitut-ing structures: A practice lens for studying technol-ogy in organizations. Organization Science, 11:404-428.

Post, J. E., Preston, L. E., & Sachs, S. 2002. Managing theextended enterprise: The new stakebolder view. Cal-ifornia Management Review, 45(1): 6-28.

Pratt, M. C , & Foreman, P. O. 2000. Classifying man-agerial responses to multiple organizational iden-tities. Academy of Management Review, 25: 18 -42.

Purdy, J. M., & Cray, B. 2009. Conflicting logics, mecba-nisms of diffusion, and multilevel djmamics inemerging institutional fields. Academy of Manage-ment Journal, 52: 355-380.

Quinn, D. P., & Jones, T. M. 1995. An agent morality viewof business policy! Academy of Management Re-view, 20: 22-42.

Ragin, C. C. 2008. Redesigning social inquiry: Fuzzysets and beyond. Chicago: University of CbicagoPress.

Ragin, C. C, Drass, K. A., & Davey, S. 2006. Fuzzy-set/qualitative comparative analysis 2.0. Tucson: De-partment of Sociology, University of Arizona.

Ragin, C. C, & Fiss, P. 2008. Net effects versus configu-rations: An empirical demonstration. In C. C. Ragin(Ed.), Redesigning social inquiry: Fuzzy sets andbeyond: 190-212. Cbicago: University of ChicagoPress.

Redding, K., & Viterna, J. S. 1999. Political demands,political opportunities: Explaining tbe differentialsuccess of left-libertarian parties. Social Forces, 78:491-510.

Riboux, B., & Ragin, C. C. (Eds.). 2008. Configurationalcomparative methods. Qualitative comparativeanalysis (QCA) and related techniques. ThousandOaks and London: Sage.

Schneider, M., Schulze-Bentrop, C, & Paunescu, M.2010. Mapping the institutional capital of high-techfirms: A fuzzy-set Emalysis of capitalist variety andexport performance. Journal of International Busi-ness Studies, 41: 246-266.

Scott, W. R., & Meyer, J. W. 1983. The organization ofsocietal sectors. In J. W. Meyer & W. R. Scott (Eds.),Organizational environments: Ritual and rational-ity: 129-154. Newbury Park, CA: Sage.

Sherif, M. 1966. In common predicament: Social psy-chology of intergroup conflict and cooperation.Boston: Houghton-Mifflin.

Sivaramakrishnan, K. 1994. Information asymmetry, par-ticipation, and long-term contracts. ManagementScience, 40: 1228-1244.

Sonenshein, S. 2006. The role of construction, intuition,and justification in responding to ethical issues atwork: The sensemaking-intuition model. Academyof Management Review, 32: 1022-1040.

Tilcsik, A. 2010. From ritual to reality: Demography,ideology, and decoupling in a post-communist gov-ernment agency. Academy of Mcmagement Journal,53: 1474-1498.


Thornton, P. H., & Ocasio, W. 2008. Institutional logics.In R. Greenwood, C. Oliver, S. K. Andersen, & R.Suddaby (Eds.), Handbook of organizational insti-tutionalism: 99-129. Newbury Park, CA: Sage.

Uzzi, B., & Gillespie, J. 1999. Corporate social capital andthe cost of financial capital: An embeddedness ap-proach. In R. Leenders & S. Gabbay (Eds.), Corporatesocial capital and liability: 446-459. Boston: Klu-wer.

Weaver, G. R., Treviño, L. K., & Cochran, P. L. 1999.Integrated and decoupled corporate social perfor-mance: Management commitments, external pres-sures, and corporate ethics practices. Academy ofManagement Journal, 42: 539-552.

Westphal, J. D., & Zajac, E. J. 1994. Substance and sym-bolism in GEOs' long-term incentive plans. Admin-istrative Science Quarterly, 39: 367-390.

Westphal, J. D., & Zajac, E. J. 1998. The symbolic man-agement of stockholders: Gorporate governance re-forms and shareholder reactions. AdministrativeScience Quarterly, 43: 127-153.

Westphal, J. D., & Zajac, E. J. 2001. Decoupling policyfrom practice: The case of stock repurchase pro-grcmis. Administrative Science Quarterly, 46: 202-228.

Winter, S. G., & Szulanski, G. 2001. Replication as strat-egy. Organization Science, 12: 730-743.

World Bank. 2003. Strengthening implementation ofcorporate social responsibility in global supplychains. Washington, DC: World Bank Group.

Yoshikawa, T., Tsui-Auch, L. S., & McGuire, J. 2007.Gorporate governance reform as institutional inno-vation: The case of Japan. Organization Science, 18:885-897.

Donal Crilly ([email protected]) is an assistant profes-sor of strategy and entrepreneurship at the London Busi-ness School. His research focuses on the role of manage-rial cognition and attention in explaining how firmsrespond to stakeholder concerns and on the applicationof linguistic analysis to understand how managers devisestrategy under uncertainty. He received his Ph.D. in strat-egy fi:om INSEAD.

Maurizio Zoilo ([email protected]) is theDean's Professor of Gorporate Strategy and Sustainabilityat Bocconi University and director of the Genter for Re-search in Organization and Management (CROMA). Hereceived his Ph.D. from the Wharton School of the Uni-versity of Pennsylvania. His research focuses on thestudy of organizational evolution through the combineduse of strategic initiatives, socialization processes, andknowledge management techniques.

Morten T. Hansen ([email protected]) is aprofessor at the University of California, Berkeley(School of Information) and at INSEAD. He holds a Ph.D.from the Graduate School of Business at Stanford Uni-versity. His research has focused on intraorganizationalsocial networks, knowledge sharing, and innovation.

Copyright of Academy of Management Journal is the property of Academy of Management and its content may

not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written

permission. However, users may print, download, or email articles for individual use.

faking it or muddling through? understanding ......tributing to the data collection. we are indebted...

Documents