TOWARD A KNOWLEDGE-BASED THEORY OF THE FIRM - Webs

Knowledge-based Theory of the FirmTransferabilityThe resource-based view of the firm recognizesthe transferability of a firm's resources and capabilities as a critical determinant of their capacityto confer sustainable competitive advantage(Barney, 1986). With regard to knowledge, theissue of transferability is important, not onlybetween firms, but even more critically, withinthe firm. The management literature has een knowing how and knowing about whichis captured by distinctions between subjective vs.objective knowledge, implicit or tacit vs. explicitknowledge, personal vs. prepositional knowledge,and procedural vs. declarative knowledge. Mypurpose here is not to make fine distinctionsbetween different types of knowledge. I identifyknowing how with tacit knowledge, and knowingabout facts and theories with explicit knowledge.The critical distinction between the two lies intransferability and the mechanisms for transferacross individuals, across space, and across time.Explicit knowledge is revealed by its communication. This ease of communication is its fundamental property. Indeed information has traditionally been viewed by economists as being a publicgood—once created it can be consumed byadditional users at close to zero marginal cost.Tacit knowledge is revealed through its application. If tacit knowledge cannot be codified andcan only be observed through its application andacquired through practice, its transfer betweenpeople is slow, costly, and uncertain (Kogut andZander, 1992).Capacity for aggregationThe efficiency with which knowledge can betransferred depends, in part, upon knowledge'spotential for aggregation. Knowledge transferinvolves both transmission and receipt. Knowledge receipt has been analyzed in terms of theabsorptive capacity of the recipient (Cohen andLevinthal, 1990). At both individual and organizational levels, knowledge absorption dependsupon the recipient's ability to add new knowledgeto existing knowledge. This requires additivitybetween different elements of knowledge.duction, mainly because this is the most important and complex means of value creation.111Efficiency of knowledge aggregation is greatlyenhanced when knowledge can be expressed interms of common language. Statistics is a particularly useful language for aggregating �its efficiency in this role is greatlyenhanced through advances in information technology. Thus, information on Ford Motor Company's cash balances, its foreign currencyexposure, its inventories of spark plugs and crankshafts is readily transferred from multiplelocations within the company and aggregated ata single location. Conversely information aboutthe capabilities of Ford managers, or the quirksof individual machine tools, is idiosyncraticknowledge which cannot be aggregated at a singlelocation. Hayek (1945: 521) refers to this as'knowledge of the particular circumstances oftime and place,' and Jensen and Meckling (1992)as 'specific knowledge.' As these authors haveshown, and as we shall explore later in the paper,the ability to transfer and aggregate knowledgeis a key determinant of the optimal location ofdecision-making authority within the firm.AppropriabiiityAppropriability refers to the ability of the ownerof a resource to receive a return equal to the valuecreated by that resource (Teece, 1987; Levin etal., 1987). Knowledge is a resource which issubject to uniquely complex problems of appropriability. Tacit knowledge is not directly appropriable because it cannot be directly transferred: itcan be appropriated only through its applicationto productive activity. Explicit knowledge suffersfrom two key problems of appropriability: first,as a public or nonrivalrous good, any one whoacquires it can resell without losing it (Arrow,1984); second, the mere act of marketing knowledge makes it available to potential buyers(Arrow, 1971: 152). Thus, except for patents andcopyrights where knowledge owners are protectedby legally established property rights, knowledgeis generally inappropriable by means of markettransactions. Lack of clear property rights resultsin ambiguity over the ownership of knowledge.While most explicit knowledge and all tacitknowledge is stored within individuals, much ofthis knowledge is created within the firm and isfirm specific. This creates difficulties over theallocation of the returns to knowledge and achiev-

112R. M. Granting optimal investment(Rosen, 1991).innewknowledgeSpecialization in knowledge acquisitionFundamental to Simon's principle of boundedrationality is recognition that the human brainhas limited capacity to acquire, store and processknowledge. The result is that efficiency in knowledge production (by which I mean the creationof new knowledge, the acquisition of existingknowledge, and storage of knowledge) requiresthat individuals specialize in particular areas ofknowledge. This implies that experts are (almost)invariably specialists, while jacks-of-all-trades aremasters-of-none.The knowledge requirements of productionProduction involves the transformation of inputsinto outputs. Fundamental to a knowledge-basedtheory of the firm is the assumption that thecritical input in production and primary sourceof value is knowledge. Indeed, if we were toresurrect a single-factor theory of value in thetradition of the classical economists' labor theoryof value or the French Physiocrats land-basedtheory of value, then the only defensible approachwould be a knowledge-based theory of value,on the grounds that all human productivity isknowledge dependent, and machines are simplyembodiments of knowledge.THE EXISTENCE OF THE FIRMThe above precepts establish a rationale for theexistence of firms. Following Demsetz (1991:171-175), the existence of the firm representsa response to a fundamental asymmetry in theeconomics of knowledge: knowledge acquisitionrequires greater specialization than is needed forits utilization. Hence, production requires thecoordinated efforts of individual specialists whopossess many different types of knowledge. Yetmarkets are unable to undertake this coordinatingrole because of their failure in the face of (a)the immobility of tacit knowledge and (b) therisk of expropriation of explicit knowledge bythe potential buyer. Hence, firms exist as institutions for producing goods and services becausethey can create conditions under which multipleindividuals can integrate their specialist knowledge. These conditions include propinquity and'low-powered' incentives designed to foster coordination between individual specialists whichavoid the problems of opportunism associatedwith the 'high-powered' incentives directly relatedto knowledge transactions.A possible solution to the inability of marketsto contract over transfers of tacit knowledge isto contract over units of workers' time. But evenif units of labor time are suitable proxies for thesupply of tacit knowledge, so long as productionrequires the complex integration of multiple typesof knowledge within a system of team production,then Rosen (1991) shows that markets mustestablish an incredibly complex wage structurewhich sets a separate wage rate for every worker's interaction with every other worker.-'Note that this view of the role of the firm asa knowledge-integrating institution is somewhatdifferent from that emphasized in the literature.Most research into organizational learning (Levittand March, 1988; Huber, 1991) and the knowledge-based view of the firm (Spender, 1989;Nonaka, 1991, 1994) focuses upon the acquisitionand creation of organizational knowledge. Thus,Spender (1989: 185) defines 'the organizationas, in essence, a body of knowledge about theorganization's circumstances, resources, causalmechanisms, objectives, attitudes, policies, andso forth.' My approach is distinguished by twoassumptions: first, that knowledge creation is anindividual activity; second, that the primary roleof firms is in the application of existing knowledge to the production of goods and services.This dispensing with the concept of organizational knowledge in favor of emphasizing the roleof the individual in creating and storing knowledge is consistent with Simon's observation that:'All learning takes place inside individual humanheads; an organization learns in only two ways:(a) by the learning of its members, or (b) byingesting new members who have knowledge theorganization didn't previously have' (Simon,1991: 125). More importantly, however, is thedesire to understand the organizational processesthrough which firms access and utilize the knowl-' On a simple production process involving n workers, whereeach worker interacts separately with each other worker, atotal of {n -n)/2 wage rates must be established (Rosen,1991: 78-81).

Knowledge-based Theory of the Firmedge possessed by their members. The dangerinherent in the concept of organizational knowledge is that, by viewing the organization as theentity which creates, stores and deploys knowledge, the organizational processes through whichindividuals engage in these activities may beobscured. Thus, March views organizations asstoring 'knowledge in their procedures, norms,rules, and forms. They accumulate such knowledge over time learning from their members'(March, 1991: 73). This learning process involves'encoding inferences from history into routinesthat guide behavior. The generic term routinesincludes the forms, rules, procedures, conventions,strategies, and technologies around which organizations are constructed and through which theyoperate' (Levitt and March, 1988: 320). Takingthe organization as the unit of analysis not onlyruns the risk of reification, but, by defining rules,procedures, conventions, and norms as knowledgefails to direct attention to the mechanisms throughwhich this 'organizational knowledge' is createdthrough the interactions of individuals, and offerslittle guidance as to how managers can influencethese processes.Unlike Spender (1992), who analyzes the dualrole of firms in knowledge generation and knowledge application, my emphasis is on the firm asan institution for knowledge application. This isnot to deny the importance of organizational context in knowledge creation. If production creationrequires the integration of each person's knowledge with that of others, even if knowledge acquisition is individualistic, the firm provides necessary incentives and direction. If knowledge isspecific to a particular team production process,then knowledge creation cannot be separated fromknowledge application—both occur within a common organizational context. Thus, if the membersof Manchester United soccer team have complementary skills, then they need to be tiedtogether by long-term relationships in order toachieve the investment in team-based skillsrequired to maximize team performance. Marketcontracts are unlikely to achieve the stability oflong-term relationships and are likely to give riseto all the problems of opportunism that transactions cost economics predicts are a consequenceof small numbers and transaction-specific investments.This rationale for the existence of the firmmay be criticized as being a special case of the113Coase/Williamson transaction cost theory of thefirm. Firms exist because they are able to avoidthe costs associated with market transactions; theknowledge-based view simply focuses upon thecosts associated with a specific type oftransaction—those involving knowledge. Certainly, the above analysis draws upon some familiar concepts of market failure. However, thekey distinction is emphasis upon the firm as anorganization for managing team production ratherthan an institution for managing transactions. Incommon with the arguments of Ghoshal andMoran (1996), the central advantage of firms inthe production process is not simply an avoidanceof the transactions costs associated with marketexchange, but their 'unique advantages for governing certain types of economic activities froma logic that is very different from that of a market(Ghoshal and Moran, 1996: 13). Integrating theknowledge of many different individuals in theprocess of producing goods and services is sucha logic. To develop this argument further, theseprocesses for integrating knowledge need to bespecified more clearly.COORDINATION WITHIN THE FIRMThe assumptions that there are gains from specialization in knowledge acquisition and storage, andthat production requires the input of a wide rangeof specialized knowledge, restates a premisewhich, either explicitly or implicitly, is fundamental to all theories of the firm. Without benefitsfrom specialization there is no need for organizations comprising multiple individuals. Given theefficiency gains of specialization, the fundamentaltask of organization is to coordinate the effortsof many specialists. Although widely addressed,organization theory lacks a rigorous integrated,well-developed and widely agreed theory of coordination.Comparative neglect of the mechanismsthrough which individuals integrate their productive activities refiects organization theory'spreoccupation, not with coordination per se, butwith the problems of cooperation which arisefrom reconciling and subordinating the disparategoals of organizational members. Thus, Lawrenceand Lorsch (1967), building upon the ideas ofMarch and Simon (1958) and Selznick (1948),viewed coordination as the resolution of intraor-

114R. M. Grantganizational goal conflict, while the institutionaleconomics literature has been dominated by theproblems of the divergence of employee andowner goals causing problems of agency (Jensenand Meckling, 1976), shirking (Leibenstein, 1966;Alchian and Demsetz, 1972) and opportunism(Williamson, 1975).Consistent with this emphasis, organizationtheory's focus upon hierarchy as the basic structure for organizing complex social activity hasconcentrated upon authority relations wherecooperation is achieved through verticallyimposed bureaucratic processes. Later writersidentified multiple mechanisms for coordinationwithin organizations. Ouchi (1979) identifiedthree types of coordination mechanism: marketmechanisms, bureaucratic mechanisms, and clanmechanisms.The knowledge-based literature has, so far, hadonly limited impact on the analysis of coordination. Research into organizational learning andmanagement of technology has explored the transfer and diffusion of knowledge within organizations (e.g., Kay, 1979; Levitt and March, 1988;Boisot, 1995), but has made only limited progressin addressing the fact that, if most of the knowledge relevant to production is tacit, then transferof knowledge between organizational members isexceptionally difficult. Nonaka (1994) emphasizesthe conversion of tacit into explicit knowledge(and vice versa), while Brown and Duguid (1991)stress the role of communities-of-practice in providing common structure and meaning for thetransfer of experience.But transferring knowledge is not an efficientapproach to integrating knowledge. If productionrequires the integration of many people's specialist knowledge, the key to efficiency is to achieveeffective integration while minimizing knowledgetransfer through cross-learning by organizationalmembers. If Grant and Spender wish to write ajoint paper together, efficiency is maximized notby Grant learning everything that Spender knows(and vice versa), but by establishing a modeof interaction such that Grant's knowledge ofeconomics is integrated with Spender's knowledge of philosophy, psychology and technology,while minimizing the time spent transferringknowledge between them.Viewing the firm's primary task as integratingthe specialized knowledge of multiple individualssuggests that, even with goal congruence, achiev-ing effective coordination is problematic fororganizations. The literature addressing integration across specialized organizational units hasviewed coordination as dependent upon thecharacteristicsof the processtechnologydeployed. Thus, Thompson identified three typesof interdependence, pooled, sequential, andreciprocal, to which Van de Ven, Delbecq, andKoenig (1976) added a fourth, team interdependence. The type of interdependence within a taskdetermines the mode of coordination deployed.Pooled interdependence calls for coordination byrules, sequential interdependence can be effectively coordinated by plans, reciprocal interdependence is associated with mutual adjustment, whileteam interdependence requires group coordination, through scheduled and unscheduled meetings (Thompson, 1967; Van de Ven et al, 1976).A knowledge-based view of the firm encourages us to perceive interdependence as an elementof organizational design and the subject of managerial choice rather than exogenously driven bythe prevailing production technology. The generalissue is devising mechanisms for integrating individuals' specialized knowledge. While processtechnology defines the technical aspects of production and the types of specialized knowledgerequired for the process, the division of tasksbetween individuals and departments and thespecification of the interfaces between them lieswithin the domain of organizational design.Integrating the literature on formal and explicitcoordination mechanisms with that on informaland implicit coordination processes, and relatingthis to characteristics and role of knowledge,points to four mechanisms for integrating specialized knowledge:1. Rules and directives. 'Impersonal': approachesto coordination involve 'plans, schedules, forecasts, rules, policies and procedures, and standardized information and communication systems' (Van de Ven et al., 1976: 323). Rulesmay be viewed as standards which regulatethe interactions between individuals. Thus, insociety at large, rules in the form of etiquette,politeness and social norms are essential tofacilitating human interaction. The efficiencyof these mechanisms in achieving coordinationextends beyond their ability to minimize communication (Galbraith, 1973). As recognizedby Demsetz (1991) direction is a 'low cost

Knowledge-based Theory of the Firmmethod of communicating between specialistsand the large number of persons who eitherare non-specialists or who are specialists inother fields' (Demsetz, 1991: 172). Such rulesare directives provide a means by which tacitknowledge can be converted into readily comprehensible explicit knowledge. Thus, it ishighly inefficient for a quality engineer toteach every production worker all that heknows about quality control. A more efficientmeans of integrating his knowledge into theproduction process is for him to establish aset of procedures and rules for quality control.2. Sequencing. Probably the simplest means bywhich individuals can integrate their specialistknowledge while minimizing communicationand continuous coordination is to organize production activities in a time-pattemed sequencesuch that each specialist's input occurs independently through being assigned a separatetime slot. Thompson viewed sequential interdependence as technologically determined.Certainly, the characteristics of the product, itsphysical inputs, and its production technologystrongly influence the potential for sequencing:a product comprised of multiple componentsfacilitates sequencing much more than a commodity produced by continuous processes.However, in most production activities thereis discretion over the extent of sequencing.For example, new product design can be fullysequential, overlapping sequences, or concurrent (Nonaka, 1990; Clark and Fujimoto,1992).3. Routines. An organizational routine is a 'relatively complex pattern of behavior . triggeredby a relatively small number of initiating signals or choices and functioning as recognizableunit in a relatively automatic fashion' (Winter,1986: 165). While routines may be simplesequences, their interesting feature is theirability to support complex patterns of interactions between individuals in the absence ofrules, directives, or even significant verbalcommunication. To this extent, routinesembody Thompson's notion of coordinationby mutual adjustment. There are two maindimensions to this complexity. First, routinesare capable of supporting a high level ofsimultaneity of individuals' performance oftheir particular tasks—examples include navigation of a ship (Hutchins, 1991), surgical115operating teams and auto racing pit crews(Grant, 1996), and the operations of fast foodrestaurants (Leidner, 1993). Second, routinescan permit highly varied sequences of interaction. While Nelson and Winter (1982) andGersick and Hackman (1990) have emphasizedthe automatic nature of routines, Pentland andRueter (1994) have shown that a routine canbe a varied repertoire of responses in whichindividuals' moves are patterned as 'grammarsof action.'4. Group problem solving and decision making.While all the above mechanisms seekefficiency of integration through avoiding thecosts of communication and learning, sometasks may require more personal and communication-intensive forms of integration. Galbraith (1973) points to the need for 'impersonal' coordination through rules and plans tobe supplemented by 'personal' and 'group'coordination modes, the last taking the formof meetings. Reliance upon ismsincreases with task complexity (Perrow, 1967)and task uncertainty (Galbraith, 1973: Van deVen et al., 1976). Hutchins (1991) documentsthe switch from routine-mode to group problem-solving mode in a crisis. The main contribution of the knowledge-based view to thisdiscussion is recognition of the high costs ofconsensus decision making given the difficulties of communicating tacit knowledge. Hence,efficiency in organizations tends to be associated with maximizing the use of rules, routinesand other integrati

Organizational theory addresses aspects of the Key words: nation knowledge; theory of the firm; coordi-firm ignored by neoclassical economics. Dispos-ing of the notion of the firm as a singular decision taker and recognizing the firm as a complex organization encompassing multiple individuals, organization theory analyzes the internal structure