سیستم های تولید انعطاف پذیر و ساختار داخلی شرکت

We investigate the relationship between the internal structure of the firm and the extent of flexibility of its technology. We demonstrate that increased vertical separation within the firm as implied, for instance, by subcontracting or by additional vertical layers of management yields investment in a more flexible technology. In contrast, increased horizontal separation as implied by lack of cooperation among different (horizontal) divisions within the firm has ambiguous implications on flexibility. When divisions of the organization confront significantly different levels of uncertainty, horizontal separation enhances technological flexibility. Otherwise, when the extent of uncertainty confronting different divisions is comparable it is cooperation that yields greater flexibility of the technology. We show that the attributes of the technology selected by a given firm may depend upon the internal structure of its competitor if those attributes can be observed by the competing firm. In particular, the firm chooses a less flexible technology if its competitor is vertically separated rather than integrated.

The face of manufacturing has been changing significantly in the late twentieth century. Flexible manufacturing systems, robotization, computer integrated manufacturing and ‘just in time’ systems have been considered essential in guaranteeing success in the market place. The shift to flexible manufacturing has been accompanied by new organizational strategies and workforce management policies. Greater reliance on teamwork and cooperation among different functions within the firm as well as increased utilization of independent suppliers and subcontractors have been key ingredients of the organizational changes. These two adjustments of the organizational form are consistent with the flexible, general purpose type of investments that modern manufacturing entails. The flexibility, which facilitates frequent adjustments of the production line and, an extended product mix, requires greater coordination among traditionally separate functions of design, engineering and marketing, thus yielding the increased reliance on teams (see Milgrom and Roberts, 1990). The general purpose type of investment reduces the extent of asset specificity in relationships with independent contractors. Since investment incentives are improved, as a result vertical separation becomes a more attractive option, as explained in the transaction cost literature (Klein et al., 1978 and Tirole, 1986; or Williamson, 1986). In the present paper we reconsider the relationship between the flexibility of the technology and the internal structure of the firm by focusing more narrowly on the inherent motive for flexibility in the production process. Essentially, firms seek flexibility since they face uncertainties in their environment and wish to respond more smoothly and less expensively to fluctuations in consumers’ tastes or in cost conditions. Given that the firm chooses to adjust its production to different realizations of the state of the world, the extent of flexibility of the technology is directly related to the extent of ex-post variability of its production decision. We evaluate how this variability depends upon the internal structure of the firm as reflected by a choice between vertical integration or separation and the extent of cooperation the firm fosters among its different (horizontal) divisions. The underlying assumption behind our analysis is the existence of asymmetry of information between the owner of the firm and her managers (or her outside independent contractors) as well as among managers of different divisions within the firm. The choice among different organizational structures affects the extent of informational asymmetries and, as a result, the nature of the agency costs incurred by the firm. Since such agency costs influence the variability of the firm’s production decision, the desired level of technological flexibility depends upon organizational structure. To incorporate environmental uncertainty in our model we assume that two separate activities in the production process are subject to random shocks the sum of which determines the cost of production. When the firm is vertically integrated the owner can observe the random shocks herself and condition the production decision on such observations. In contrast, with vertical separation it is only the manager in charge of a given activity that can observe the random shock affecting this activity. With vertical separation, the owner can choose between an organizational structure that fosters cooperation between the managers and one that prohibits such cooperation between them. In the former case, the two managers act as a single agent and in the latter, managers make their decisions independently. We investigate how each of the above three organizational forms (i.e., vertical integration, vertical separation+cooperation, and vertical separation with no cooperation) affect the variability of the production decision and as a result, the incentives of the owner to invest in flexible manufacturing techniques. It is noteworthy that in addition to serving as a model of the internal organization of the firm our model can also be interpreted as explaining industrial organization.1 With such an alternative interpretation our analysis can explain the implications of vertical or horizontal mergers within an industry upon the incentives of the merged entity to invest in flexible manufacturing techniques. Our model does not include explicit assumptions concerning cost complementaries between flexibility and certain organizational forms. In contrast, the existing literature on modern manufacturing (Milgrom and Roberts, 1990) focuses primarily on such complementarities between coordination, diversification of the product line and short term relationships with suppliers. Specifically, when the cost of coordinating the functions of separate units declines so does the cost of extending the product mix or negotiating contracts with suppliers. In our model, it is the extent of asymmetry of information of different organizational forms that determines the extent of variability in the firm’s production decision and the implied desired level of flexibility of its technology. To abstract from issues related to economies of scope we focus only on scale and not product-mix flexibility. Hence, if the owner invests in a more flexible technology she incurs lower adjustment costs when varying the quantities of output of the single product she produces. Even though we assume that the choice of organizational structure of a given firm is always observable to its competitor, we distinguish between two different regimes related to the observability of the attributes of the technology that are selected by the firm. In the Unobservable Regime a given firm cannot observe the attributes of the technology selected by its competitor and in the Observable Regime such information is available among competing firms. While in the Unobservable Regime the choice of technology is determined solely by the extent of variability in the production decision of the firm, in the Observable Regime additional strategic considerations yield distorted investment decisions and ex-post productive inefficiencies. At the Cournot equilibrium this additional strategic effect yields reduced investment in flexibility and expanded production, aimed at inducing the competitor to cut back on its output. We demonstrate that, in general, vertical separation of the firm yields greater production variability and, therefore, enhanced investment in flexibility. Specifically, restricting consideration to subgame perfect equilibria, we assume that a firm who remains vertically separated experiences a greater degree of asymmetry of information between the owner and her managers than if it were vertically integrated. This increased asymmetry induces greater ex-post variability in its production, thus requiring larger investments in enhanced flexibility. The vertically separated firm makes the larger investments irrespective of whether or not it is the only one to remain vertically separated in the industry. In contrast, the implications of horizontal separation between divisions upon flexibility are ambiguous and depend upon the extent to which uncertainties vary across different divisions. With a focus on economies of scope, Milgrom and Roberts predicted a definite association between teamwork and product-mix flexibility. Enhanced cooperation and communication among managers of different divisions is implied by the more diversified product mix that intensified (product-mix) flexibility supports. In our model, where volume flexibility is considered, cooperation between different divisions need not always be consistent with a more flexible manufacturing technique. Since cooperation among divisions may have an ambiguous effect on the extent of asymmetry of information in the organization, the extent of variability of the firm’s production decision does not necessarily increase with cooperation. We find that under the Observable Regime the owner’s choice of technology does not only depend upon the internal structure of her own firm but that of the competing firm as well. Specifically, the vertical separation of the rival yields reduced investment in flexibility. Both under the Observable and Unobservable Regimes asymmetric equilibria may arise even though firms are assumed to be a-priori identical. At such equilibria, firms may choose different forms of organizational structure as well as different degrees of flexibility of their technologies. Although the literature on technological flexibility is rather large and covers many fields from engineering, operations research, human resource management, and production management, to economics and game theory, most papers on the subject take a decision theoretical perspective (see, for instance, Gerwin, 1993, Jones and Ostroy, 1984 and Milgrom and Roberts, 1990; or the pioneering work of Stigler, 1939). The papers that do address the strategic implications of flexibility in competitive markets do not investigate the interaction between flexibility and the internal organization of the firm. Vives, 1989 and Vives, 1993 studies, for instance, the relationship between the incentives of competing firms to invest in more flexible technologies and the extent of either prior or posterior uncertainty that is confronting them. He demonstrates that with privately informed oligopolistic firms there need not be a tradeoff between the value of strategic commitment and the degree of flexibility of the selected technology. Kulatilaka and Marks (1988) demonstrate the strategic value of technological flexibility in negotiating wages and working conditions with labor unions. Röller and Tombak (1990) study the market forces that drive industries into adopting flexible manufacturing systems. They find that increased product differentiation and larger markets are likely to encourage flexibility in the technology. Eaton and Schmitt (1994) demonstrate that flexible manufacturing techniques have significant implications on market structure. Similar to Milgrom and Roberts (1990), they focus attention on product-mix flexibility and demonstrate that the existence of economies of scope and cost complementaries promote market concentration through preemption and mergers. Spencer and Brander (1992) derive a tradeoff between flexibility and commitment that has some resemblance to the strategic effect we derive in our model under the Observable Regime. In their model, however, the tradeoff arises since increased flexibility yields delays in investment and output decisions which can be adversely exploited by competitors. Boyer and Moreaux (1997) demonstrate the possible existence of asymmetric equilibria with one firm investing in a flexible manufacturing system while its competitor choosing the inflexible technology. In contrast to the asymmetry that we obtain, it is not different organizational structures that yield the divergence in the firm’s behavior, but the fact that the benefits from flexibility decline when the competitor invests in a more flexible technology as well. None of the above papers addresses the linkage we find in the present analysis between the extent of asymmetry of information within an organization and the extent of variability in production. The internal structure of the firm determines the extent of asymmetry of information between the owner of a firm and her managers. Any structure that increases the asymmetry results also in greater variability of the firm’s production decision and necessitates, therefore, a more flexible manufacturing technique. The rest of this article is organized as follows. In the next section (Section 2) we describe the basic assumptions of our duopoly model. In Section 3 we evaluate, as a benchmark the vertically integrated case and conduct a comparison of the Observable and Unobservable Regimes in this context. In Section 4 we derive the subgame perfect equilibria of the entire game under the Unobservable regime and in Section 5 we do the same under the Observable regime. Section 6 concludes the paper.

We have investigated the implication of the internal structure of the firm upon the extent of flexibility of its technology. We found that increased vertical separation within the firm, as implied, for instance, by subcontracting or by additional vertical layers of management yields investment in a more flexible technology. In contrast, the implication of increased horizontal separation upon technological flexibility is ambiguous and depends upon the relative uncertainties confronting various activities within the firm. Adopting a team approach under which cooperation among various activities and/or departments is encouraged yields greater technological flexibility only if the various activities face comparable levels of uncertainty. Otherwise, if the extent of uncertainty confronting activities is significantly different it is lack of cooperation and the discouragement of teamwork which yields greater technological flexibility. Even though we have obtained the results for a very simple, binary specification of the distribution functions, the effect of internal structure upon the extent of variability in production can be extended to continuous distribution functions as well. The fact that intensified asymmetry of information increases the variability of production holds for more general distribution functions. Incorporating competitive forces in the model allowed us to investigate whether the characteristics of the technology chosen by one firm depend upon the technological choices or the organizational structure of its competitor. We found that such a dependency exists only when the attributes of the technology of the firm are observable to its competitor. The choice of technology, in this case, is dictated by both strategic and productive efficiency considerations. In essence, while in the absence of observability flexibility is solely determined by the extent of variability of the environmental uncertainty,13 with observable technologies each firm evaluates also the implications of its choice upon the production decision of its competitor. Such additional strategic considerations yield reduced flexibility of the technologies selected by the firms. In addition, each firm selects a less flexible technology when its competitor is vertically separated rather than vertically integrated. The importance of the observability of the technology is illustrated also in Boyer and Moreaux (1997) and in Spencer and Brander (1992). Both papers derive the strategic value of commitment to a certain technology under the assumption that this commitment is fully observable to the competitor. We have demonstrated that firms may choose different attributes of their technologies and different forms of organizational structures even though they face identical demand and cost conditions. In particular, while one firm may choose to be vertically integrated its competitor may remain vertically separated. As a result, the former type of firm chooses a less flexible technology than the latter type of firm. A similar asymmetry was also obtained in Boyer and Moreaux. In our model, the decision concerning the internal structure of the firm precedes the choice of technology. If this assumption is reversed so that technology is chosen first our results remain unaffected in the Unobservable regime. In the Observable regime an additional strategic effect determines the attributes of the technology. Those attributes are chosen then to affect not only output decisions but the rival’s choice of internal organization as well. Even though our model has been confined to analyzing flexibility with respect to rates of production and not with respect to the mix of products produced it can be extended to allow for multiproduct flexibility as well. Incorporating economies of scope and cost complementarities is essential if such an extension is pursued. Milgrom and Roberts (1990) have demonstrated, for instance, that the existence of cost complementarities should dictate cooperation among different divisions whenever more flexible manufacturing techniques are adopted by the firm. Hence the ambiguity that we obtain concerning the relationship between flexibility and cooperation may disappear with cost complementarities and a diversified product mix. In our model the only potential agency problem that arises due to vertical separation stems from the existence of asymmetric information between the owner and her managers. It is assumed, however, that the owner can perfectly enforce any desired quality level of the product that she wishes. Relaxing this latter assumption adds a moral hazard problem to the analysis and may yield more complex forms of contracting between the owner and her managers. In Taylor and Wiggins (1997) it has been demonstrated, for instance, that in order to enforce quality the firm may have to inspect the output produced by its supplier (manager) or threaten to discontinue its relationship with him in case of substandard performance. While all of our results are obtained for specific demand and cost functions as well as very simple distributional properties of the uncertainties, our qualitative predictions are likely to remain unaffected in a more general framework. In essence, in the Unobservable Regime flexibility is determined by the extent of variability in the output decision of the firm. Any organizational structure that increases this variance implies a more flexible manufacturing technique. Either enhanced vertical separation or enhanced horizontal separation when uncertainties affecting different divisions are significantly different both lead to such increased fluctuations in the production. In the Observable Regime an additional strategic effect, aimed at manipulating the rival’s choice of output, may distort the choice of the technology and yield inefficiencies in production.