Inter-Temporal Economies of Scope, Organizational Modularity, and the Dynamics of Diversification Constance E

Strategic Management Journal Strat. Mgmt. J., 25: 1217–1232 (2004) Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/smj.427

INTER-TEMPORAL ECONOMIES OF SCOPE, ORGANIZATIONAL MODULARITY, AND THE DYNAMICS OF DIVERSIFICATION CONSTANCE E. HELFAT1* and KATHLEEN M. EISENHARDT2 1 Tuck School of Business, Dartmouth College, Hanover, New Hampshire, U.S.A. 2 Department of Management Science and Engineering, Stanford University, Stanford, California, U.S.A.

The question of whether corporations add value beyond that created by individual businesses has engendered much debate in recent years. Some of this debate has focused on the pros and cons of related vs. unrelated diversification. A standard explanation of the benefits of related diversification has to do with the ability to obtain intra-temporal economies of scope from contemporaneous sharing of resources by related businesses within the firm. In contrast, this paper deals with inter-temporal economies of scope that firms achieve by redeploying resources and capabilities between related businesses over time, as firms exit some markets while entering others. The transfer of resources due to market exit distinguishes our treatment of inter-temporal economies of scope from standard intra-temporal economies of scope. In addition, these inter- temporal economies can benefit from a decentralized and modular organizational structure. This ability to obtain inter-temporal economies of scope via organizational modularity and recombination suggests that corporations do not necessarily need a high degree of coordination between business units in order to benefit from a strategy of related diversification. Copyright  2004 John Wiley & Sons, Ltd.

INTRODUCTION utilization of, or reduction in, excess resources within the firm to obtain economies of scope. Such The question of whether corporations add value economies of scope (e.g., Bailey and Friedlander, beyond that created by individual businesses has 1982; Panzar and Willig, 1981) derive from the engendered much debate in recent years (e.g., contemporaneous sharing of tangible or intangi- Rumelt, 1991; Bowman and Helfat, 2001; Brush, ble assets in the production of multiple products, Bromiley, and Hendrickx, 1999; Martin and Eisen- resulting in lower joint costs of production per unit hardt, 2003a). An important aspect of this debate of output. The resulting intra-temporal economies deals with the pros and cons of unrelated vs. of scope arguably comprise an important benefit related diversification. Although the evidence is of the related diversified corporation. not clear cut, many studies have suggested that In contrast, this paper is concerned with inter- firms may benefit from related diversification in temporal economies of scope derived from the particular (see Montgomery, 1994; Palich, Car- redeployment of firm resources between businesses dinal, and Miller, 2000). A standard explanation over time, as firms exit some product-markets of the benefits of related diversification involves while entering others. Inter-temporal economies of scope have implications for the evolution of Keywords: economies of scope; diversification; organiza- related diversified corporations over time, partic- tion form; modularity; market entry ularly in markets where technologies and demand *Correspondence to: Constance E. Helfat, Tuck School of Busi- ness, Dartmouth College, 100 Tuck Hall, Hanover, NH 03755, are in flux. As these sorts of markets emerge, U.S.A. E-mail: [email protected] grow, split, combine with other markets, and

Copyright  2004 John Wiley & Sons, Ltd. Received 11 January 2002 Final revision received 14 May 2004 1218 C. E. Helfat and K. M. Eisenhardt mature, managers must make frequent strategic The analysis begins with a brief review that decisions about whether and when to enter or exit explains how the standard intra-temporal logic of product-markets. In what follows, we use the term economies of scope justifies related diversifica- ‘dynamic’ to denote these sorts of fast-paced mar- tion. Then we turn to the dynamics of continuing kets (see also Eisenhardt and Martin, 2000). related diversification over time. We present our Inter-temporal economies of scope in dynamic arguments regarding inter-temporal economies of markets can benefit from an organizational form scope and formalize the concept. The analysis then that consists of a modular, decentralized organiza- turns to the organizational ramifications for firms tional structure along with processes for recombin- seeking to benefit from inter-temporal economies ing businesses (product-market domains in which of scope as part of a strategy of related diversifica- the firm competes) among modular organizational tion through entry and exit. We also provide a more units. This sort of repeated recombination of orga- detailed example of a firm that continually pur- nizational units to match changing business oppor- sues related diversification over time, and achieves tunities is a form of ‘patching’ (Brown and Eisen- inter-temporal as well as intra-temporal economies hardt, 1998; Siggelkow, 2002) that results in an of scope while employing a highly decentralized evolving path of related diversification through and modular organizational structure. Then we time. As organizational units shift their product- discuss the implications of this empirical exam- market responsibilities over time (Galunic and ple for related diversification, organizational form, and the dynamics of diversification. The conclud- Eisenhardt, 1996, 2001), the units co-evolve with ing section draws broader implications regarding the products they produce and the associated market entry and exit, the adaptation of firms to organizational, technical, and market knowledge changing markets and technologies, and empirical (Helfat and Raubitschek, 2000). Additionally, the research on related diversification. ability to obtain inter-temporal economies of scope within a decentralized organization contrasts with standard organizational prescriptions for obtaining ECONOMIES OF SCOPE AND intra-temporal economies of scope (see, for exam- DIVERSIFICATION ple, Hill, Hitt, and Hoskisson, 1992). Corporations Penrose (1995, first edition published in 1959) lays do not necessarily need a high degree of coordi- out the foundation for the analysis of firm growth nation between business units in order to benefit over time via continued diversification. Penrose from a corporate strategy of related diversification identifies two forms of what we now term ‘related’ over time. diversification: (1) entry into new product-markets Our work extends the literature in several ways. based on the firm’s existing resources; (2) intro- First, we explain the concept of inter-temporal duction of new products in a firm’s existing economies of scope, formalize the concept in a market (Penrose, 1995: 110). Rumelt (1974: 29) parsimonious manner, and explain the link to value states that: ‘[b]usinesses are related to one another creation through related diversification. Secondly, when a common skill, resource, market or purpose we indicate an organizational form, involving applies to each.’ As these early works make clear, recombination of modular organizational units, that the sharing of firm resources between businesses is useful in obtaining inter-temporal economies of underpins much of the logic of related diversifica- scope in dynamic markets. Although not the only tion. The concept of economies of scope, initially route to economies of scope in dynamic markets developed in a separate literature, formalizes the (see also Martin and Eisenhardt, 2003b), modu- logic that links shared resources to related diversi- lar recombination is an important organizational fication. form that has been overlooked in the literature Economies of scope (Panzar and Willig, 1981) on related diversification. As part of this analysis for two products at a point in time are defined as1 of modular recombination, we also suggest ways + in which intra-temporal economies of scope are C(Y1,Y2)

Copyright  2004 John Wiley & Sons, Ltd. Strat. Mgmt. J., 25: 1217–1232 (2004) Inter-temporal Economies of Scope 1219 where C = total costs of production, Y1 = output participates, the economies of scope just described of product 1, and Y2 = output of product 2. are static in the sense that they are intra-temporal. Equation 1 states that the total cost of producing That is, once two businesses merge, the combined Y1 and Y2 together is less than the combined cost firm achieves intra-temporal economies of scope of producing each product separately. As Bailey by sharing resources contemporaneously (intra- and Friedlander (1982) explain, this reduction in temporally) between businesses and cutting redun- costs due to joint production can arise for several dant costs. Similarly, subsequent to market entry reasons. These reasons include: (1) separate prod- through internal growth, a firm achieves economies ucts that naturally arise from a shared input, such of scope by contemporaneous sharing of what pre- as wool and mutton produced from sheep; (2) the viously were excess resources. presence of a fixed factor of production (e.g., a Although firms that diversify into related busi- manufacturing plant or distribution channel) that is nesses may benefit from intra-temporal economies not fully utilized in production of a single product; of scope, Teece (1980) points out that such joint (3) economies of networking from joint production ownership of resources is efficient only when of networked products (e.g., use of an airline hub the transaction costs of separate ownership (due to facilitate transfer of passengers from one air- to costs of contracting and opportunism) can be line city-pair market to another); (4) reuse of an reduced through internal organization (Williamson, input in more than one product (e.g., journal arti- 1975). Since internalization of transactions also cle abstracts reused in multiple indexes of articles); entails costs, diversification based on economies (5) sharing of intangible assets between products of scope should occur only if the costs of internal (e.g., research and development that supports mul- organization are lower than the transactions costs tiple products). of using the market (i.e., production in separate All of these reasons can help to explain diversifi- organizations). In the remainder of this analysis, cation from one product (Y1) into another product we assume that the transaction cost criterion for (Y2) due to inputs (resources) shared in produc- internal organization of joint production has been tion of the products. In general, diversification can satisfied. take place via merger and acquisition, or via inter- The logic of economies of scope formalizes nal growth. Diversification through merger and the benefits of related diversification in terms of acquisition often generates economies of scope by cost advantages. These benefits from economies of allowing firms to share a fixed factor of produc- scope can also be formulated in terms of demand- tion and cut redundant costs. For example, when side benefits related to outputs (products and ser- a company that produces cheese acquires a com- vices) rather than costs. For example, when firms pany that produces crackers, the two businesses use excess resources to diversify into another mar- can share the fixed factor associated with grocery ket, the firm generates greater revenues per unit store distribution and sales. As a result, the dis- of input. This is logically equivalent to Equation 1 tribution and sales resources of the acquired firm for economies of scope, wherein the firm obtains become redundant and their costs can be cut. lower costs per unit of output by spreading the cost Firms also can diversify through internal growth of a set of inputs over a greater number of units and obtain economies of scope. A cracker manu- of output. Our previous examples of economies of facturer can enter the cheese business and share scope from internal growth reflect precisely this grocery store distribution and sales between the logic. Thus, intra-temporal economies of scope two businesses. This example of diversification reflect both demand-side revenue enhancements through internal growth translates the shared cost from greater output and cost reductions from logic of economies of scope into a motivation for shared inputs. market entry. An excess resource of some type provides the opportunity to reduce unit costs by diver- sifying and sharing that resource with another busi- DYNAMICS OF RELATED ness. Economies of scope also can justify simul- DIVERSIFICATION taneous rather than sequential entry into related markets. Although firms can obtain intra-temporal econo- Although by definition a diversification move mies of scope after diversification into related mar- implies a change in the businesses in which a firm kets, the standard treatment of economies of scope

Copyright  2004 John Wiley & Sons, Ltd. Strat. Mgmt. J., 25: 1217–1232 (2004) 1220 C. E. Helfat and K. M. Eisenhardt does not deal directly with the dynamic aspects of could still be interpreted as a series of intra- related diversification. Furthermore, most empir- temporal scope economies obtained through a ical studies of diversification examine a cross- series of related diversification moves over time.2 sectional snapshot of company business portfolios Penrose (1995) herself refers to such ‘economies of (for a review, see Montgomery, 1994; Ramanu- production’ obtained after expansion takes place. jam and Varadarajan, 1989; Palich et al., 2000). Our analysis focuses on a different aspect of Only occasionally do studies analyze diversifica- related diversification over time that involves inter- tion moves from an existing product-market into temporal redeployment of resources within a firm another domain (examples include Montgomery in response to changing market conditions. Inter- and Hariharan, 1991; Mitchell, 1989; Helfat, 1997; temporal economies of scope derive from entry Silverman, 1999). In reality, managers often diver- into new product-markets in conjunction with par- sify their firms through a series of moves that occur tial or complete exit from old product-markets. In over an extended time period. this situation, a new business partially or com- Diversification that unfolds over time can be pletely replaces an old business and the firm modeled as a series of single diversification moves shifts resources from the old to the new busi- that result in standard intra-temporal economies ness. This redeployment of resources differs from of scope. Such an approach, however, does not the contemporaneous sharing of resources between explain where the excess resources that support businesses that underlies standard intra-temporal continuing diversification come from. For this, economies of scope. we require additional theory, for which the main The phenomenon of inter-temporal resource source is Penrose (1995). transfer in response to changing market conditions Penrose (1995) provides three general explana- receives little attention from Penrose (1995). She tions for the continuing production and/or posses- does, however, briefly observe that firms continue sion of excess resources, which we briefly note to expand even when demand for their original here. The first explanation involves economies of products falls or disappears. Wernerfelt (1984) and scope from indivisible assets. The second closely Anand and Singh (1997) also observe that when a related explanation involves assets that are spe- market collapses, a firm still might be able to make cialized to particular tasks. Due to their generally good use of a resource from the collapsed market smaller size, young firms cannot efficiently utilize by using the resource in another market. Penrose these indivisible and specialized assets. As firms (1995) further notes in passing that as markets expand to a size large enough to fully employ the mature, they may become relatively less profitable indivisible assets and to utilize assets in their most compared with new investment opportunities that productive specialized use, growth and diversifica- firms could undertake using their resources. tion eventually cease. Kim and Kogut’s (1996) study of entry by indi- Penrose’s (1995) third explanation involves effi- vidual semiconductor firms into new technical sub- ciencies gained from learning by doing that result fields has implications for these types of situations. in excess resources, which the firm can use to In their analysis, Kim and Kogut (1996) show that expand existing businesses or to move into related when entering new subfields, semiconductor firms businesses. Penrose (1995) argues that firms use built on knowledge accumulated from prior entry knowledge accumulated through previous diversi- into related markets.3 These patterns of market fication moves as a basis for subsequent diver- entry over time mirrored technological trajecto- sification. Thus, firms continuously extend their ries, where some technologies displaced others. knowledge bases over time via entry into related product-markets (Helfat and Raubitschek, 2000). 2 Rubin (1973) provides a formal mathematical model that cap- Consistent with this logic, Chang (1997) finds that tures some of Penrose’s (1995) logic. In Rubin’s (1973) model, a firms are more likely to enter industries that have firm’s resources can be used to produce current output as well as greater knowledge similarity to the industries that larger amounts of the same resource in future periods. He notes that the model can explain diversification as well as expansion they entered in the preceding time period. in the firm’s current market. The importance of indivisible assets, specialized 3 Some subfields served as platforms for expansion into other resources, and especially learning by doing helps subfields, by virtue of either network externalities of compo- nents in a linked technological system or of learning by doing us to better understand what underlies the dynam- that enabled firms to accumulate knowledge on which to base ics of diversification. The end result, however, expansion.

Thus, although Kim and Kogut (1996) do not pro- time periods: vide direct evidence, the ﬁrms might have exited markets where the technologies were displaced, C(Y1,t−1,Y2,t )

C(Y2i,t |Y1i,t−1)

INTER-TEMPORAL ECONOMIES OF 4 In an analysis of product tying and entry deterrence, Carl- SCOPE ton and Waldman (2002) also briefly refer to inter-temporal economies of scope. The authors do not define the term and use it in a narrow context where lower costs for one product To add precision to the analysis of inter-temporal in the second period have implications for entry into another resource transfers between related businesses, we product that has demand-side complementarities. 5 next modify the standard formula for economies In this set-up, producing Y1 essentially creates real options for the firm to move into some potential set of Y s in the future. of scope. We define inter-temporal economies 2 Some Y1s may give the firm more choice among potential Y2s of scope as follows, for two products and two than do other Y1s.

The analysis of inter-temporal economies of scope replace existing businesses with new businesses pertains to redeployment of resources between without changing either the number of businesses businesses over time in reaction to permanent in which it participates or the size of the firm, and changes in technologies and market demand. The still obtain inter-temporal economies of scope. analysis does not apply to a repeated pattern Inter-temporal economies apply only to busi- of resource transfer between ongoing businesses, nesses that are related to one another by skill, such as that which occurs in some seasonal busi- resource, market, or purpose (per Rumelt, 1974). nesses. For example, many ski areas redeploy their This relatedness of businesses is critical. Unless facilities and staff every summer for warm weather the business that a firm exits is related to one that it mountain activities, and then shift these resources enters, the firm cannot redeploy resources between back to the ski business in the winter. Although the two businesses. Inter-temporal economies also seasonal resource transfers conform to the formula do not accrue through merger and acquisition for inter-temporal economies of scope, they are strategies of the sort where firms buy and sell more appropriately modeled using the formula for companies without moving resources. In addition, intra-temporal economies of scope given earlier. our analysis of inter-temporal economies does not The ski area, for example, is permanently in two rely on learning-by-doing, in contrast to Penrose’s businesses and has not exited either market. Thus, (1995) emphasis on learning-by-doing to explain we can model the company as sharing its resources diversification over time. Learning-by-doing, as between two businesses over a time period of one noted earlier, could be viewed as leading to a series year.6 Since the company is in the same businesses of intra-temporal scope economies over time. year after year, this is properly viewed as a stable phenomenon, rather than as a series of diversifica- Adjustment costs tion moves. The key element in our analysis that differenti- Because inter-temporal economies of scope result ates inter-temporal from intra-temporal economies when one business partially or completely replaces of scope is market exit. When a firm permanently another related business, it is important to con- reduces its presence in one business and adds sider the adjustment costs of transferring resources another business, inter-temporal economies arise between businesses, including the time that it takes from resource redeployment. Although the process to redeploy the resources. The concept of adjust- of exit and entry may recur over time as markets ment costs comes from the economics literature change, in each instance of market exit combined on investment and growth. In economics, adjust- with entry the firm permanently alters the busi- ment costs stem from difficulties that firms may nesses in which it participates.7 Moreover, unlike face in adjusting their capital stock due to market in standard analyses of related diversification and imperfections that raise the costs and time period intra-temporal economies of scope, inter-temporal required for investment (Lucas, 1967). For exam- economies do not imply that the size and scope of ple, a firm that invests in specialized equipment the firm necessarily increase over time. A firm can from a small supplier may need to pay a premium in order to obtain a large order or face a time delay in receiving the equipment. If we think of diver- 6 For example, Levy and Haber (1986) analyze shifts of resources sification as an investment in a new market, then back and forth between a stable set of product lines when demand for the different products fluctuates. Their analysis, adjustment costs become relevant, even for stan- which incorporates some uncertainty about demand shifts and dard analyses of diversification that do not incorpo- includes adjustment costs of shifting resources between product rate inter-temporal economies of scope. The litera- lines, essentially is a generalization of the situation of seasonal shift in resource usage. Fernandez-Cornejo et al. (1992) also ture on diversification, however, tends not to focus estimate an empirical model that includes adjustment costs of on adjustment costs of moving between markets. shifting resources between a fixed set of product lines over time. What sort of adjustment costs might inter- 7 It is of course possible to assume away the time dimension by temporal resource redeployment involve? As a first defining the relevant time period as a long enough period, of say 10 or 20 years, such that the firm can be viewed as participat- cut, we can divide them into direct and indirect ing in all of the businesses during the ‘same’ time period. Such costs of resource transfer. Direct costs of resource a model, which includes a business that the firm has already transfer include the expense of moving people permanently exited as a component of intra-temporal (contemporaneous) economies of scope, would obscure the dynamics of and equipment between businesses. Indirect costs diversification through market exit and entry. result from disruption to existing businesses during

Copyright  2004 John Wiley & Sons, Ltd. Strat. Mgmt. J., 25: 1217–1232 (2004) Inter-temporal Economies of Scope 1223 these moves and from the amount of time involved firm may enter a new related business to which in resource transfer. Time delays result in foregone it switches the product development team (inter- revenue and therefore constitute a cost of resource temporal scope economies), and for which it shares transfer. Adjustment costs of resource transfer production and marketing with the mature business relate to what Helfat and Raubitschek (2000) call (intra-temporal scope economies). ‘integrative knowledge’ and what Mitchell and In what follows, we provide a detailed exam- Shaver (2003) term ‘integration capability.’ Both ple of a firm that operates in dynamic markets terms refer to the ability of an organization to and obtains both inter-temporal and intra-temporal absorb new businesses and to manage a variety of economies of scope. This example elaborates on different businesses on a continuing basis. Firms how excess resources arise as a result of chang- that seek inter-temporal economies of scope on a ing markets and explains the organizational mech- regular basis may develop integrative capabilities anisms used by this firm to transfer resources directed toward lowering the adjustment costs of between businesses. The organizational structure resource transfer. and processes of this firm are particularly relevant Adjustment costs would matter less if the shift to adjustment costs of resource transfer. For this from Y1 to Y2 reflected a change from one long- reason, we first provide an overview of the liter- lasting equilibrium to another. In this situation, ature on related diversification and organizational permanent cost reductions obtained from econo- form before proceeding to our company example. mies of scope would make up for one-time adjustment costs. In dynamic markets, however, product- life cycles are short and markets mature and ORGANIZATIONAL FORM AND even shrink quickly. This limits the time horizon RELATED DIVERSIFICATION during which the firm can obtain inter-temporal economies of scope between any two businesses. As a result, the firm can profit from inter-temporal The topic of diversification and organizational economies of scope only if adjustment costs do form has to do with the broad subject of ‘strategy not offset the benefits of inter-temporal resource and structure.’ As documented by Alfred Chandler transfer between businesses. The length of time (1962), many prominent, large, diversified firms in that it takes to redeploy resources becomes espe- the United States adopted a divisionalized orga- cially important. If resource transfer occurs slowly, nizational structure that contrasted with earlier the time period during which the firm can obtain structures that organized tasks according to func- inter-temporal economies of scope is shortened. tion rather than product-market. Oliver Williamson The need to minimize both the amount of time (1975) later referred to this type of organiza- that it takes to redeploy resources between busi- tion as the M-form. Within this structure, opera- nesses and the direct costs of transferring resources tional decision making and control occurred at the has implications for organizational form. For these level of divisions organized according to product- reasons, in the next section we analyze issues of markets, and strategic decision making took place organizational form. at the top of the organization. The M-form reduced In practice, diversified firms have the poten- the need for coordination between product lines tial to benefit from both inter-temporal and intra- within functions, thus economizing on the coor- temporal economies of scope. A firm might trans- dination capacity of top management. In addition, fer resources over time between some businesses delegation of operational control economized on and share resources contemporaneously between the bounded rationality of top managers, enabling other businesses. In addition, if a firm reduces but them to focus on strategic issues for the company does not completely eliminate an existing business as a whole. Using the M-form, top management while also entering another business, the firm may also could reward division managers based on the simultaneously benefit from both inter-temporal performance of their units, creating stronger (more and intra-temporal economies of scope. When a high-powered) incentives. Finally, by using divi- business matures, for example, a firm may decide sions as profit centers, top management could more to partially or completely eliminate new prod- effectively monitor divisions, reducing the poten- uct development for that business, but maintain tial for agency problems within the corporation production and marketing. At the same time, the (Williamson, 1991).

In the strong form just described, the M-form significant effect on return on assets for related does not provide a mechanism for achieving econo- diversified firms. They also find that greater mies of scope across divisions, except insofar as decentralization, less reliance on non-financial top management might transfer information from criteria, and use of division-level performance one division to another as part of its strategic incentives have a statistically significant positive advice to the divisions. As a result, in diversi- effect on return on assets for unrelated diversified fied firms where divisions operate in related mar- firms. kets, firms may need to modify the strict M- As these results suggest, and as Hill et al. (1992) form in order to achieve economies of scope point out, diversified firms may face a trade- across divisions. Accordingly, Hill et al. (1992), off with regard to their organizational arrange- in a comprehensive review of the literature, posit ments: the characteristics that promote economies that four organizational characteristics affect the of scope can harm governance economies and success of related diversification and the asso- vice versa. Recent research has suggested vari- ciated realization of economies of scope. These ous alternatives for dealing with this dilemma. characteristics have to do with the extent to One new approach is that of patching (Brown which corporations employ: (1) centralized con- and Eisenhardt, 1998), whereby top management trol over strategic and operating decisions of divi- adds, splits, combines, and removes decentralized sions; (2) coordination between divisions through and modular business units over time as markets integrating mechanisms; (3) non-financial criteria change. Because the fast pace of dynamic markets to evaluate divisional performance; (4) incentives makes it more difficult to coordinate operations and rewards tied to corporate rather than divisional across businesses from the top of the organiza- profitability. tion, patching dispenses with ongoing coordina- Hill et al. (1992) argue that centralized control tion between divisions that might achieve intra- and coordination improve the ability of related temporal economies of scope. Instead, more lim- diversified firms to share resources across divisions ited economies of scope may arise from factors in order to obtain economies of scope. Centralized such as the use of a common brand name, product control can help identify opportunities for resource development system, and accounting system across sharing and ensure that division managers seek businesses. In addition, because firms that rely on to exploit these opportunities. Coordination facil- patching are decentralized, they can reward man- itated by integrating mechanisms such as cross- agers based on division profits, reducing agency division teams can help to implement the sharing problems within the firm. of resources across divisions. Non-financial eval- Other approaches seek to benefit from cross- uation criteria, such as those that reward coordi- division economies of scope without changing nation between divisions, provide incentives for the architecture of the business portfolio of the cross-division integration to achieve economies firm. Goold and Campbell (1987), for exam- of scope. Finally, when firms use corporate-wide ple, have argued that an intermediate organiza- financial incentives, managers have an incentive tional form that they term ‘strategic control’ can to share resources with other divisions in the firm, incorporate aspects of both decentralization and which can lead to economies of scope. Hill et al. coordination. In particular, when division man- (1992) also argue that, unlike related diversified agers are rewarded based on division perfor- firms, unrelated diversifiers should reward man- mance, they should have an incentive to improve agers based on financial measures of divisional the performance of their divisions, including by performance rather than corporate performance, in obtaining useful resources from other divisions order to obtain governance economies. In addi- (Goold, Campbell, and Alexander, 1994). Mar- tion, centralization and coordination would harm tin and Eisenhardt (2003b) take a more granu- the performance of unrelated diversified firms by lar view and observe that this approach requires raising their costs without benefits from economies some way for division managers to learn about of scope. valuable resources in other divisions and to nego- Hill et al. (1992) find that greater use of tiate the details of resource sharing. They propose centralization, interdivisional integration, non- that, as part of the process of ‘co-evolving,’ multi- financial criteria, and corporate-wide financial business teams comprised of the heads of indi- incentives have a positive and statistically vidual businesses can identify collaborative links

Copyright  2004 John Wiley & Sons, Ltd. Strat. Mgmt. J., 25: 1217–1232 (2004) Inter-temporal Economies of Scope 1225 among businesses as markets evolve, and negotiate modular organizational structure, with only limited mutually beneficial ways to achieve these collabo- sharing of resources between corporate divisions rations. Thus, as markets shift, these teams enable on an ongoing basis. Inter-temporal resource trans- firms to figure out and then exploit where new fer occurs by a process of recombining organiza- intra-temporal economies of scope lie. tional units with businesses in a manner described These newer solutions focus on the benefits next.8 of decentralization and high-powered incentives found in the M-form, and then build in resource ECONOMIES OF SCOPE IN OMNI sharing across divisions in a way that does not CORPORATION preclude decentralization. Yet, firms still face a tension between decentralization and resource shar- Omni Corporation is a Fortune 100 high-tech- ing, and these solutions are likely to favor decen- nology company that participates in many busi- tralization over coordination and economies of nesses with overlapping customers and related scope. Moreover, like the older literature on diver- technologies, and has been one of the most con- sification strategy and structure, these newer solu- sistently high-performing technology corporations tions involve intra-temporal economies of scope in the world.9 During the time period of the mid that derive from contemporaneous (albeit dynamic) 1990s described here, Omni’s businesses included sharing of resources between divisions. electronic instrumentation (from voltage meters Unlike standard intra-temporal economies of to sophisticated microwave analyzers), comput- scope, achieving inter-temporal economies of ing and information technology (handheld com- scope does not create a tension between decen- puters, desktop computers, and servers), and com- tralization of businesses and coordination across puter peripherals (such as monitors and printers). businesses that firms must overcome with cre- The company was organized into several busi- ative solutions. Because inter-temporal economies ness groups, each composed of multiple divisions do not involve contemporaneous resource shar- related to one another on the basis of product lines, ing, achieving these economies does not require customer segments, and technology. Each divi- ongoing coordination and integration across division had global strategic and operational control sions. This reduced need for coordination implies over its business, and operated as a profit cen- that firms can decentralize divisions and use ter within the corporation. That is, Omni had a division-level financial performance as the basis decentralized organizational structure, where busi- for high-powered incentives and monitoring of nesses were run largely separately and managers managers. Firms do require some centralization were rewarded for the performance of their busi- of decisions regarding market entry, exit, and nesses and products (high-powered incentives). A the resulting redeployment of resources, but do division could produce more than one product, but not require centralized control of division strat- each division had relatively narrow product-market egy and operations. As a result, related diversifiers responsibilities. that pursue inter-temporal economies of scope can Omni participated in markets that often changed benefit from decentralized business units, high- rapidly as a result of evolving and new technolo- powered financial incentives, and monitoring of gies, and shifts in customer desires once new tech- managers based on division performance. In con- nologies were commercialized. To deal with these trast with other approaches to achieving economies ongoing technological and market changes, Omni of scope in related diversified firms, inter-temporal executives reallocated what they called ‘charters’ economies of scope do not conflict with gover- among its divisions. A charter was defined as nance economies to nearly the same extent. a product-market area of responsibility, such as In what follows, we provide a detailed example ‘desktop computing.’ When company executives of inter-temporal economies of scope and associated organizational arrangements in a company 8 This recombination process is a subset of the corporate ‘patch- that participates in dynamic markets. The exam- ing’ strategies discussed earlier (Brown and Eisenhardt, 1998). ple involves a company, Omni Corporation (a The literature on patching has yet to identify the link to inter- pseudonym used to preserve confidentiality), that temporal economies of scope analyzed here. 9 The facts about Omni Corporation described here draw in part pursues a corporate strategy of related diversi- on Galunic and Eisenhardt (1996, 2001), but the interpretation fication. The firm also has a decentralized and offered with regard to economies of scope is new.

Copyright  2004 John Wiley & Sons, Ltd. Strat. Mgmt. J., 25: 1217–1232 (2004) 1226 C. E. Helfat and K. M. Eisenhardt moved charters between divisions, they moved factors such as entry by competitors, which tended only product-market responsibility, not staff or to lower product prices. The divisions were no other resources (with the exception of isolated longer in a fast-growth mode, and in some situ- managerial changes). ations sales were slipping. Managers could have The charter changes in Omni consisted of tried to increase margins or sales growth in the three main types. First, new business opportunities divisions’ existing charters. Raising profitability, resulting from technological and market changes however, would have been difficult in the face led to new charters. Second, when shifts of increased competition in markets where the in the nature of markets and technologies products had become more of a commodity (and produced a poor fit between a division’s therefore, lower margin almost by definition) due capabilities and the business specified by the to maturation of the underlying technology. Man- charter, Omni executives moved the charter agers also could have laid off personnel and sold to another division. Third, businesses with off equipment. pressing difficulties precipitated temporary charter As an alternative, Omni’s method of charter reallocations. To accomplish these charter changes, recombination transferred resources within a divi- which occurred with regularity, Omni executives sion to a new product-market area of respon- used a few main approaches, as next described. sibility, which partially supplanted the declining The company executives approached the first type business from which the new business drew its of charter change, which involved new business resources.10 The new business opportunity offered opportunities, in a manner that enabled the firm to the prospect of much higher margins than did obtain inter-temporal economies of scope. For this the existing charter in a mature market. As Pen- reason, we devote particular attention to this type rose (1995) notes, when markets mature, they of charter change. We also describe the other two may become relatively less profitable compared sorts of charter changes. with new investment opportunities that firms could undertake using their resources. In Omni’s case, New business opportunities resources were redeployed by recombining divisions with business opportunities (known as char- Corporate management played an important role ters at Omni) through time. in spotting new related business opportunities In addition to the generation of new charters by for Omni in emerging markets and technologies. corporate management, there was a second source Omni’s corporate management used its knowl- of new charters: new business ideas sometimes edge of the firm as a whole to actively search for emerged from divisions within Omni, stemming and evaluate new related diversification opportuni- from ongoing operations. Omni executives, how- ties. For example, in one instance, Omni corporate ever, often considered these business ideas to be executives realized that technologies being devel- ‘new charters,’ took them away from divisions that oped in two different divisions could be merged had originated them, and (as in the previous exam- to create a new business. The way in which Omni ple) allocated them to other divisions. Generally, executives dealt with the new business opportu- the divisions that originated but lost charters for nity, however, is noteworthy. They often did not new business opportunities were growing rapidly form a new division to handle this opportunity. in their main businesses. These divisions would Instead, as occurred frequently when corporate have been hard pressed to extend their resources management identified new business opportuni- to take on new businesses as well. Although Omni ties, Omni executives allocated the new charters executives could have transferred personnel such to existing divisions that were attaining modest or as engineers and marketing specialists from else- weak performance under their current charters as where in the company to these divisions, so that a result of shrinking market growth or increased the divisions could take on new businesses, Omni competition. did not have enough managerial capacity at the top The weakened performance of the recipient divi- of the corporation to facilitate such transfers on sions had resulted in part from the normal growth and maturation of the high-technology markets 10 In many instances, Omni proactively exited businesses where in which Omni participated. As a technology sales were declining, rather than waiting until market conditions matured, margins and sales growth declined due to worsened.

Copyright  2004 John Wiley & Sons, Ltd. Strat. Mgmt. J., 25: 1217–1232 (2004) Inter-temporal Economies of Scope 1227 a regular basis. In addition, top management fre- ‘give’ the charter to a division that had skills that quently preferred to keep the divisions that inaugu- better fit with the business. For example, instead rated new charters tightly focused on their current of trying to undertake the time-consuming and businesses. potentially tortuous process of making a division The recombining of divisions and charters just whose competencies lay in R&D become a cost- described was possible only because the com- driven unit with superior manufacturing capabili- pany participated in related businesses. Without the ties, Omni executives moved the charter to a unit relatedness of businesses, so that excess resources that already had the necessary capabilities. in one division could be productively applied to The divisions that received the charters, how- a new business (charter), corporate management ever, were not necessarily those within Omni that would have found it more difficult to reallocate had the skills that best fit the charter. Instead, the resources within a division to a new business charters went to divisions that, while they had opportunity that did not necessarily arise from relevant capabilities, also had excess resources. within that division. Galunic and Eisenhardt (2001) characterize the Omni’s charter changes in order to pursue receiving divisions as somewhat successful per- new business opportunities created inter-temporal formers that were constrained by modest charters economies of scope, based on redeployment of (i.e., of limited actual and potential market size). resources from a business with slowing market As a result, these divisions had excess and related growth, or even sales declines, to a new related resources that could be applied to the additional business. In terms of the earlier formula for inter- charter they received. For example, in one case, temporal economies of scope, Y2 denotes the new the charter involved a complementary product in business opportunity and Y1 denotes the prior the same industry as the recipient division’s cur- business of the division that received the new rent business. In addition, corporate management charter, where Y2 partially (but not fully) replaced worked with the losing divisions to find new uses Y1. Excess resources that developed in period t − 1 for their now excess resources in the form of new for Y1 were transferred to Y2 in the subsequent time charters that fit better with the divisions’ compe- period, as the firm reduced its commitment to the tencies. market for Y1. These charter changes due to mismatches between product-markets and division capabilities Mismatched capabilities and markets yield the standard sort of intra-temporal economies of scope, in the following sense. When a receiving A second type of charter change within Omni division obtains an additional charter because resulted from shifts in the nature of product- its current charter has limited market size, the markets and technologies, which produced a mis- division already has excess resources available to match between the capabilities of a division and share contemporaneously with the new business. the basis of competition in the business. For exam- These excess resources might have arisen as a ple, as often occurs in high-technology industries result of prior learning by doing (Penrose, 1995), as markets mature, the competitive basis of the for example. The charter change, however, does industry for one Omni division was shifting from not involve exit from a product-market or the R&D, where the division’s competencies lay, to redeployment of resources from an exited business manufacturing efficiencies. Similarly, for another to another business that creates inter-temporal division, the basis of competition was shifting from economies of scope. the ability to serve the military customer to the commercial customer. Unexpected orphans Given the dynamic nature of the markets in which Omni competed, such mismatches occurred The third type of charter change at Omni involved frequently (as did the new business opportunities ‘orphaned’ businesses. These were often smaller previously discussed). As mismatches emerged, businesses that did not fit well within their current divisions often encroached upon the charters of divisions and posed pressing problems. Corporate other divisions. Corporate management would then management often quickly dealt with these prob- intervene to formally allocate the charter either lems by moving the charter to another division to the originating division, or more frequently to (termed ‘foster homes’) on a temporary basis until

Copyright  2004 John Wiley & Sons, Ltd. Strat. Mgmt. J., 25: 1217–1232 (2004) 1228 C. E. Helfat and K. M. Eisenhardt a permanent solution could be found. One sort of group. The personnel in the receiving divisions temporary charter change occurred when divisions could also take on new businesses because Omni that were pursuing new charters were hindered by had some degree of standardization among divi- responsibility for their old businesses in a way that sions, including a common accounting system, a was not anticipated. Omni executives usually pre- common brand, and a common framework for ferred to move the old businesses to ‘foster home’ the product development process. For example, divisions so as not to hold back the new grow- products under development in one division could ing businesses. A second sort of temporary charter be moved to another division, and the receiving change involved situations where the scale of oper- division would have a clear understanding of the ations had shrunk to the point where the businesses progress and status of the development project. As were difficult and/or costly to manage on a stand- a result, these standardized systems and processes alone basis. In both types of charter change, the also lowered the adjustment costs of resource rede- problems also had escalated to the point where ployment associated with charter changes. they needed immediate attention that could not The charter changes within Omni also reveal wait for corporate management to fully diagnose ways that related diversified firms can achieve the sources of the problems and devise a permanent both inter-temporal and intra-temporal economies solution. of scope, as firms redeploy some resources among The foster home divisions that received tempo- businesses over time and share other resources rary charters were well-run, solid performers with between businesses contemporaneously. The first stable markets. In the parlance of economies of type of charter change in Omni, involving new scope, while these divisions did not have large business opportunities combined with partial and amounts of excess resources, they were not run- occasionally complete exits of old businesses, ning completely flat out and could share their resulted in inter-temporal economies of scope. current resources to accommodate a small char- These inter-temporal economies may also have ter addition, resulting in intra-temporal economies been accompanied by intra-temporal economies of scope. of scope in cases where some resources (e.g., a distribution channel) were shared contemporaneously between the new business and any remaining DISCUSSION portion of the old business. Moreover, because systems and processes were shared contemporane- Omni’s approach to diversification has implica- ously between the divisions, they produced intra- tions for several features of our analysis of inter- temporal economies of scope as well. The other temporal economies of scope. As next explained, two types of charter changes also involved contem- Omni’s approach adds to our understanding of: poraneous resource sharing between businesses, (1) value creation through related diversification, resulting in intra-temporal economies of scope that (2) an organizational form for achieving value are more dynamic in character than the standard through related diversification that runs counter to treatment of shared resources across divisions. the prevailing wisdom, and (3) how a combination More generally, firms can achieve both inter- of market entries, exits, and resource recombina- temporal and intra-temporal economies of scope tions supports an evolving set of related businesses not only through entry into related product- over time. markets, but also from introduction of new generations of products in the same market. Related diversification Resources that are shared contemporaneously The charter change processes within Omni reveal across two generations of a product that important ways in which relatedness of businesses coexist yield intra-temporal economies of scope, is linked to achieving corporate value. Omni is and resources that are redeployed one from a related diversified firm and the relatedness of product generation to another yield inter-temporal Omni’s businesses was critical to the firm’s ability economies of scope. to transfer charters among businesses over time. Personnel in the receiving divisions could take on Organizational form the new businesses because the knowledge bases Omni’s approach highlights the importance of an of the divisions were related within each business appropriate organizational form in order to create

Copyright  2004 John Wiley & Sons, Ltd. Strat. Mgmt. J., 25: 1217–1232 (2004) Inter-temporal Economies of Scope 1229 value from related diversification. In allocating type of organizational arrangement, however, tends resources to businesses (charters), Omni used a to entail higher coordination costs that erode some decentralized and modular organizational structure of the benefits of decentralization. Although firms that runs counter to the view that related diversified that compete in multiple geographic and product firms ought to benefit from centralized coordina- markets sometimes utilize a matrix type of arrange- tion and control. Omni’s objective financial crite- ment, the complexity of this approach adds to the ria and incentives based on division performance, costs of day-to-day operations. A matrix organi- as well as the lack of extensive integration and zation also makes it more difficult to obtain the coordination mechanisms across divisions, also are governance economies that Omni achieves through inconsistent with the traditional view (e.g., Hill high-powered evaluation and incentive systems et al., 1992). Although Omni did centralize deci- tied to division performance. sion making regarding charter transfers, most if In addition to the benefits of Omni’s orga- not all of the decisions on business strategy and nizational form for inter-temporal economies of related operating matters were made at the divi- scope, Omni appears to have successfully applied sion level. While the particular types of charter this form of organization to obtain intra-temporal changes just described are specific to Omni, other economies of scope via reallocation of existing related diversified firms have employed a simi- charters. The benefits of applying this modular approach of recombining organizational units lar, decentralized organizational structure to exist- with business charters, although perhaps not as ing businesses may come about because Omni extensively. Examples include Johnson and John- participates in dynamic markets, where the firm son, 3M, Microsoft, Dell, Adobe, and Symantec. must continually readjust the match between its Thus, in contrast to more standard prescriptions, resources and changing market needs. it appears that successful related diversified firms Overall, the Omni example suggests that firms can be decentralized and modular.11 can effectively employ modular recombination as Omni’s approach to inter-temporal economies a powerful organizational form that contrasts with of scope may work because its combination of the standard prescriptions for related diversified a decentralized, modular structure and recombina- firms. Particularly in dynamic markets, related tion processes helps to minimize the adjustment diversified firms may be able to achieve economies costs and amount of time for resource transfer of scope along with (rather than instead of) gov- required to obtain inter-temporal economies of ernance economies. A decentralized organization scope in dynamic markets. It may be faster to structure, low levels of coordination between divi- transfer charters between divisions than to trans- sions, and incentives and monitoring tied to divi- fer resources across divisions as markets emerge, sion performance can be used in a manner that converge, segment, mature, and decline. Espe- enables firms to obtain both inter-temporal and cially when businesses are related, a division that intra-temporal economies of scope. receives a new charter can probably come up to speed more rapidly and efficiently than when extensive resources must be transferred between Dynamics of diversification divisions. In addition, by keeping divisions intact, Omni avoids disrupting the social relationships Application of the concept of inter-temporal econo- that enable the smooth functioning of routines and mies of scope to charter changes within Omni resources within divisions. brings out several important features of the dynam- Of course, Omni’s approach may not be the ics of diversification. Omni combined exit from only one that related diversified firms might use some product-markets with entry into others, which to achieve inter-temporal economies of scope. An in turn produced inter-temporal economies of scope alternative to Omni’s approach might involve a as the company redeployed resources between matrix type of organization, where personnel and related businesses over time. This process created plant and equipment in a particular functional area exactly the sort of evolving path of dynamic related could be reassigned to a new business. A matrix diversification documented by Kim and Kogut (1996), Chang (1997), and Helfat and Raubitschek 11 It is worth noting that Omni produced superior financial (2000). Omni repeatedly used its existing knowl- performance for several decades using the approach described. edge base as a platform for expansion into new

Copyright  2004 John Wiley & Sons, Ltd. Strat. Mgmt. J., 25: 1217–1232 (2004) 1230 C. E. Helfat and K. M. Eisenhardt related businesses, even though the expansion fre- empirical research on related diversification, as quently took place within another division of the next explained. company. That is, charters for new business oppor- With regard to market entry, inter-temporal tunities developed in one division often went to economies of scope provide an additional explana- other divisions that had related resources that could tion for the empirical regularity of success through be productively redeployed to the new businesses. diversified market entry. In a well-known study, The originating divisions understood that they Dunne, Roberts, and Samuelson (1988) found that were stretched too thin to undertake the new busi- diversified entrants performed better than other nesses. The originating divisions also understood entrants. As Helfat and Lieberman (2002) explain that if and when their current businesses declined, in a comprehensive analysis of the relationship they would gain a new charter, possibly devel- between firm resources and market entry, diver- oped from elsewhere within the company. Thus, sified entrants tend to meet with greater success the process of building new knowledge and related for reasons specifically due to the nature of their diversification based on this knowledge continued. pre-entry resources. In line with this reasoning, our In addition to allocating charters for new busi- analysis suggests that inter-temporal economies of ness opportunities, Omni reallocated existing busi- scope may contribute to the success of diversified ness charters between divisions. Although these entrants through resource redeployment between reallocations were not new diversification moves, businesses, because inter-temporal economies of they supported the related diversification posture scope lower entry costs for related diversifiers rel- of the company by seeking to improve the perfor- ative to newcomers. mance of the related businesses within the cor- Inter-temporal economies of scope also point to poration through frequent readjustment of divi- the advantages of timely market exit. In redeploy- sional capabilities with businesses. Taken as a ing resources, firms may benefit when they exit whole, Omni’s approach of combining product- markets with declining opportunities in order to market entry, exit, and business reshuffling sup- take advantage of new opportunities in other mar- ported an evolving set of related diversified busi- kets (Penrose, 1995). The ability to obtain inter- nesses within the firm. temporal economies of scope from simultaneous exit and entry provides an incentive for related diversified firms to more quickly exit declining CONCLUSION businesses. As Sull (1999) observes, failure to exit from a declining business in a timely fash- This paper asks: what value does the corporation ion can degrade the performance of the entire firm. add beyond what markets and individual busi- Firms that proactively exit declining markets while nesses can accomplish effectively? We began by at the same time successfully entering new mar- observing that the value of the corporation beyond kets are what Tushman and O’Reilly (1997) term the sum of its businesses is in debate. This debate ‘ambidextrous organizations,’ and what Brown and focuses in part on the value of related diversifi- Eisenhardt (1998) refer to as operating on ‘the edge cation and attendant economies of scope. In this of chaos.’ paper, we have analyzed more fully and with This process of market entry combined with greater organizational depth how related diversi- exit enables firms to adapt to changing mar- fied firms can obtain economies of scope, particu- kets and technologies as firms alter the compo- larly in dynamic markets. Our discussion of Omni sition of their businesses over time. For exam- identified how related diversification can create ple, product-market opportunities often shift dur- value through inter-temporal and non-traditional ing the life cycle of products and technologies. intra-temporal economies of scope, how modular As products and technologies mature and even recombination can be used to achieve that value in decline, firms may be left with resources that a manner that does not entail a trade-off between can be reapplied to new related business oppor- economies of scope and governance economies, tunities to produce inter-temporal economies of and how an evolving set of related businesses scope. Moreover, a decentralized, modular orga- results over time. The analysis also has implica- nizational structure together with recombination tions for market entry and exit, the adaptation of processes can enable managers of a related diversi- firms to changing markets and technologies, and fied firm to relatively quickly match organizational

Copyright  2004 John Wiley & Sons, Ltd. Strat. Mgmt. J., 25: 1217–1232 (2004) Inter-temporal Economies of Scope 1231 units with new business opportunities. Resources Dess, Rich D’Aveni, Isaac Fox, Vijay Govindara- from mature, declining, or defunct businesses can jan, Bob Hansen, Anne Marie Knott, Nancy Mar- be reallocated quickly to more promising production, Margie Peteraf, Dan Raff, Lori Rosenkopf, market opportunities. Thus, a modular organiza- Nicolaj Siggelkow, Gabriel Szulanski, Alva Tay- tional structure and recombination process can lor, Sid Winter, two anonymous reviewers, and enhance adaptation to changing markets and tech- seminar participants at the Max Planck Institute nologies (Brown and Eisenhardt, 1998), including for Evolutionary Systems, the University of Min- the creation of new product-market domains.12 nesota, the University of Pennsylvania, the Uni- Finally, our analysis has implications for empir- versity of Sussex, and the University of Toronto. ical research on diversification. First and foremost, studies of diversification may understate the benefits of related diversification because research gen- REFERENCES erally has not incorporated the possibility of inter- temporal economies of scope. Evaluating inter- Anand J, Singh H. 1997. Asset redeployment, acquisi- temporal scope economies empirically requires tions, and corporate strategy in declining industries. longitudinal data as well as a different research Strategic Management Journal, Summer Special Issue design that incorporates exit from some businesses 18: 99–118. Bailey EE, Friedlander AF. 1982. Market structure and entry into other related businesses, in order to and multiproduct industries. Journal of Economic capture economies of scope over time. In addition, Literature 20(3): 1024–1048. future research might test whether redeployment Bowman EH, Helfat CE. 2001. Does corporate strategy of resources within firms as part of a process of matter? Strategic Management Journal 22(1): 1–23. combined market exit and entry improves firm per- Brown SL, Eisenhardt KM. 1998. Competing on the formance relative to market exit unaccompanied Edge. Harvard Business School Press: Boston, MA. Brush TH, Bromiley P, Hendrickx M. 1999. The relative by entry into related product-markets, and rela- influence of industry and corporation on business seg- tive to the performance of start-up companies in ment performance: an alternative estimate. Strategic new markets. Research also could test whether Management Journal 20(6): 519–548. firms that seek inter-temporal economies of scope Capron LW, Mitchell W, Swaminathan A. 2001. Asset by transferring resources among businesses over divestiture following horizontal acquisitions: a dynamic view. Strategic Management Journal 22(9): time achieve better performance when they use 817–844. a modular, decentralized organization. Entry and Carlton DW, Waldman M. 2002. The strategic use of exit information can be derived from sources such tying to preserve and create market power in the Compustat business segment database and the evolving industries. RAND Journal of Economics Trinet plant-level database. Both of these databases 33(2): 194–220. Chandler AD Jr. 1962. Strategy and Structure: Chapters contain performance information as well. These in the History of the American Industrial Enterprise. data could also be supplemented with surveys of MIT Press: Cambridge, MA. companies in order to collect more detailed infor- Chandler AD Jr. 1990. Scale and Scope: The Dynamics mation on resource profiles, resource transfer, and of Industrial Capitalism. Belknap Press of Harvard organizational form. And these are but a few of University Press: Cambridge, MA. Chang SJ. 1997. An evolutionary perspective on the possibilities for future research based on the diversification and corporate restructuring: entry, exit, concept of inter-temporal economies of scope as and economic performance during 1981–89. Strategic part of a related diversification strategy. Management Journal 17(8): 587–611. Dunne T, Roberts MJ, Samuelson L. 1988. Patterns of firm entry and exit in U.S. manufacturing industries. RAND Journal of Economics 19: 495–515. ACKNOWLEDGEMENTS Eisenhardt KM, Martin JA. 2000. Dynamic capabilities: what are they? Strategic Management Journal This paper benefited from comments and sug- 21(10–11): 1105–1121. gestions from Phil Anderson, Betsy Bailey, Joel Fernandez-Cornejo J, Gempesaw CM, Elterich JG, Ste- Baum, Guido Buisendorf, Ashish Chopra, Greg fanou SE. 1992. Dynamic measures of scope and scale economies: an application to German agriculture. American Journal of Agricultural Economics 74(2): 12 See also Capron, Mitchell, and Swaminathan (2001) on the 329–342. role or mergers and acquisitions in leading to resource recombi- Galunic DC, Eisenhardt KM. 1996. The evolution of nation and the evolution of firms through time. intracorporate domains: divisional charter losses

in high-technology, multidivisional corporations. Montgomery CA. 1994. Corporate diversification. Jour- Organization Science 7(3): 255–282. nal of Economic Perspectives 8(3): 163–178. Galunic DC, Eisenhardt KM. 2001. Architectural innova- Montgomery CA, Hariharan S. 1991. Diversified expan- tion and modular corporate forms. Academy of Man- sion by large established firms. Journal of Economic agement Journal, 44(6): 1229–1249. Behavior and Organization 15: 71–89. Goold M, Campbell A. 1987. Strategies and Styles: Palich LE, Cardinal LB, Miller CC. 2000. Curvilinearity The Role of the Centre in Managing Diversified in the diversification–performance linkage: an exam- Corporations. Basil Blackwell: Oxford. ination of over three decades of research. Strategic Goold M, Campbell A, Alexander M. 1994. Corporate- Management Journal 21(2): 155–174. Level Strategy: Creating Value in the Multibusiness Panzar JC, Willig RD. 1981. Economies of scope. Company. Wiley: New York. American Economic Review 71(2): 268–272. Helfat CE. 1997. Know-how and asset complementarity Penrose ET. 1995. The Theory of the Growth of the Firm and dynamic capability accumulation: the case of (3rd edn). Oxford University Press: New York. R&D. Strategic Management Journal 18(5): 339–360. Ramanujam V, Varadarajan P. 1989. Research on corpo- Helfat CE, Lieberman MB. 2002. The birth of capabili- rate diversification: a synthesis. Strategic Management ties: market entry and the importance of pre-history. Journal 10(6): 523–551. Industrial and Corporate Change 11(4): 725–760. Rubin PH. 1973. The expansion of firms. Journal of Helfat CE, Raubitschek RS. 2000. Product sequencing: Political Economy 81: 936–949. co-evolution of knowledge, capabilities and products. Rumelt RP. 1974. Strategy, Structure, and Economic Strategic Management Journal 21(10–11): 961–980. Performance. Harvard Business School Press: Boston, Hill CWL, Hitt MA, Hoskisson RE. 1992. Cooperative MA. versus competitive structures in related and unrelated Rumelt RP. 1991. How much does industry matter? diversified firms. Organization Science 3(4): 501–521. Strategic Management Journal 12(3): 167–185. Kim D, Kogut B. 1996. Technological platforms and Siggelkow N. 2002. Evolution toward fit. Administrative diversification. Organization Science 7(3): 282–301. Science Quarterly 47: 125–159. Levy DT, Haber LJ. 1986. An advantage of the Silverman BS. 1999. Technological resources and multiproduct firm: the transferability of firm- the direction of corporate diversification: toward specific capital. Journal of Economic Behavior and an integration of the resource-based view and Organization 7(3): 291–302. transaction cost economics. Management Science Lucas RE. 1967. Adjustment costs and the theory of 45(8): 1109–1124. supply. Journal of Political Economy 75(4): 321–334. Sull DM. 1999. The dynamics of standing still: Firestone Martin JA, Eisenhardt KM. 2003a. Cross-business syn- Tire & Rubber and the radial revolution. Business ergy: sources, processes, and the capture of corporate History Review 73(3): 430–464. value. Kelleher Center Working Paper Series (Num- Teece DJ. 1980. Economies of scope and the scope of ber 3): 1–62. University of Texas at Austin: Austin, the enterprise. Journal of Economic Behavior and TX. Organization 1(3): 223–247. Martin JA, Eisenhardt KM. 2003b. Cross-business col- Tushman ML, O’Reilly CA III. 1997. Winning Through laboration: recombination, modularity, and the multi- Innovation: A Practical Guide to Leading Organiza- business team. Kelleher Center Working Paper Series tional Change and Renewal. Harvard Business School (Number 2): 1–107. University of Texas at Austin: Press: Boston, MA. Austin, TX. Wernerfelt B. 1984. A resource-based view of the firm. Mitchell W. 1989. Whether and when? Probability and Strategic Management Journal 5(2): 171–180. timing of incumbent and newcomer market share and Williamson OE. 1975. Markets and Hierarchies: Analysis survival when specialized assets retain their value. and Antitrust Implications. Free Press: New York. Administrative Science Quarterly 34: 208–230. Williamson OE. 1991. Comparative economic organiza- Mitchell W, Shaver JM. 2003. Who buys what? How tion: the analysis of discrete structural alternatives. integration capability affects acquisition incidence and Administrative Science Quarterly 36: 269–296. target choice. Strategic Organization 1(2): 171–201.