Organizing For Synergies: Allocating Control to Manage the Coordination-Incentives Tradeo

Transcription

1 Organizing For Synergies: Allocating Control to Manage the Coordination-Incentives Tradeo Wouter Dessein, Luis Garicano, and Robert Gertner Graduate School of Business The University of Chicago March 3, 2005 Abstract Merging companies attempt to realize synergies by sharing resources to exploit scale economies in functional areas such as R&D, manufacturing, or sales. If it were possible to do everything else the same, such mergers would always be value creating. However, in order to exploit the synergy, the rm may need to reallocate some decision-making authority in the relevant function and soften overall incentives. We develop a model of organizational structure in which communication is strategic and in which control is allocated to either functional or product managers. We characterize the optimal choice of organizational structures as a function of the value of synergies, the importance of incentives and the private information about, and value of, local adaptation. We use our analysis to obtain some insights in some speci c real world applications. 1 Introduction 1.1 General Introduction Organizations exist to coordinate complementary activities in the presence of specialization. Organizational structure, through its impact on the information available to decision-makers and their incentives, exhibit tradeo s between implementing synergies, adapting to local conditions, and inducing e ort. We develop a model to capture the tradeo s that guide organizational and incentive structure in the presence of coordination problems due to potential synergies. The model allows us to explore the mechanisms by which these tradeo s operate, develop predictions about when di erent organizational structures are optimal and provide a framework for thinking about other issues relating to organizational structure. These tradeo s can be critical in determining how an organization can realize synergies among di erent business units. There are countless examples of rms that fail to implement organizational strategies that allow them to realize the potential gains from a merger. Many of the most spectacular failed mergers involve the inability to achieve the synergies that 1

2 motivated the deal. The recent merger between AOL and Time Warner is a particularly prominent example of what appears to be quite common. 1 The merging parties claimed an important source of increased value from the merger would be synergies from selling advertising packages that included all media encompassed by the merged company s divisions. However, centralized ad-selling was thwarted by divisional advertising executives who felt they could get better deals than the shared revenue from centralized sales. An outside advertising executive was quoted by the Wall Street Journal, stating, "[t]he individual operations at AOL Time Warner have no interest in working with each other and no one in management has the power to make them work with each other." AOL Time Warner could have chosen to provide more authority to the centralized advertising unit, but this too is not without cost and signi cant peril. Taking authority away from business units over such an important source of revenue could reduce the sensitivity of decisions to local information, reduce the coordination among the di erent activities of a business unit, and blunt incentives. 2 The more general question is one that arises in any theoretical discussion of mergers. Typically, it is possible to identify some likely source of potential synergy in mergers of related companies by taking advantage of scale economies and sharing resources in some functional area, be it R&D, manufacturing, distribution, brand coverage, or sales. A simple (or simplistic) justi cation for the merger is that the two rms can do everything as before except the narrow combination of activities needed to exploit the synergy. The standard (economist s) response to this is that there are costs associated with expanding the scale of the rm and then there is some hand-waving about managerial diseconomies of scale that derive from increased bureaucracy and limited spans of control, and perhaps less nancial market discipline if both merging companies are public. These answers are mostly imprecise and they are not very satisfying. They do not explain why bureaucracy increases if most everything is the way it was before the merger. Perhaps more important, they do not tell us when these costs are relatively important and when they are not. Thus, we cannot make predictions about when mergers will be e ciency-enhancing because we can only speak to the details of one side of the tradeo. This lack of speci city about the costs of organizational scale is also a problem in many discussions of the theory of the rm that involve tradeo s between improved coordination and bureaucracy costs. The problem of organizing to achieve synergies is not unique to mergers; all companies with related lines of business must decide which activities to centralize, how to allocate control rights over complementary decisions, how to share relevant information, and how to create incentives for e ective coordination and e cient operations. In Section 4.5 we discuss two particularly notable cases where the tradeo s we highlight play a particularly strong role: the attempts to reorganize the FBI after the rst WTC bombing in 1993, and the reorganization of Suchard in Europe after the push towards the creation of a European single market in 1 See Rose, Angwin, and Peers (2002). 2 The anecdotal evidence of failed synergy implementation is also supported by the broader empirical literature on merger performance in corporate nance. See Andrade, Mitchell, and Sta ord (2001). 2

3 Introduction to the model We develop a model to capture the tradeo s that guide organizational structure in the presence of local information, coordination problems, e ort choice, and potential synergies. We model the following simpli ed setting. There are two independent agents 3 engaged in an economic activity. For concreteness, say they are each watchmakers who produce and market handmade watches. Customers with varying preferences arrive at their shops, and the watchmakers apply their e orts to producing watches adapted to local demand. In manufacturing the watch, they both require some component, say the watch movement, that is produced by a third party. 4 one that is optimal for the demands of its client base. Each watchmaker may demand a di erent movement, However, there are potential cost savings from producing a standard watch movement, common to both watchmakers. The magnitude of the savings are not known at the time the organization is set up, but will be known eventually by the component supplier. Producing a standard movement also has a cost, which is that the resulting watch will be less adapted to the needs of the individual watchmakers. The magnitude of these costs are also unknown at rst, and become private information of the individual watchmakers. The problem the parties have to solve is to determine the organizational structure that will be most likely to decide rightly whether they should standardize component production. An important glitch is that, realistically, all three managers must exert e ort, and e ort is unobservable. Thus, the compensation of both watchmakers and the movement manufacturer must be linked to their performance. However, and this is key, motivating managers by linking compensation to performance biases them away from the common objectives, as it makes them care about their own output, thereby making the communication process and synergy implementation decision strategic. Thus the movement manufacturer that has a stake in low costs will emphasize, in any communication with the watchmakers, the importance of building standardized components; the watchmakers in turn will tend to emphasize the importance of local adaptation. Similarly, control over the decision will show the same biases. Organizational structure optimally trades o the communication/implementation e ciency with the value of high-powered incentives for e ort. More generally, our model aims to capture a situation in which production involves two product managers, who each supply a product or service, and a functional manager who may be able to implement synergies between the activities of the two product managers. After organizing their relationships, the parties obtain private information: the functional manager 3 We will often refer to these agents as managers although in the most direct interpretation of our model, the agents need not have any employees. 4 For the purpose of the motivating example, one can think of the three parties as independent rms or as part of a single organization. We think that the better interpretation of our model is that of a single organization because enforcement of ex post authority may require common ownership. 3

4 learns about the value of synergies, the value of doing things in a standardized way across units; product managers learn about the value of adaptation to local circumstances, the value of doing things in an individualized way, or equivalently the costs of restricting adaptation by implementing the synergy. Determining optimally whether or not the synergies should be implemented requires that managers communicate. Control rights over the decision must be allocated in a way that encourages both communication and e cient implementation given the information. Moreover, managers must be motivated to work hard, and this requires that their compensation be linked to their performance. When the incentives of individual managers are su ciently strong, the interests of functional manager and product managers are directly in con ict, and no credible communication can take place. Thus stronger individual incentives lead to higher e ort, but ine cient synergy implementation decisions. The organization must deal with these tradeo s. The tools that it can use are the allocation of authority over synergy implementation and incentive pay. We analyze three possible allocations of authority. We rst study ex-ante decisions, where managers do not communicate and the organization chooses either to always implement or never implement synergies. We then consider two types of organizations that try to aggregate the available information: (1) functional authority, where the control rights over implementation decisions reside with the functional manager; and (2) product authority, where the control rights reside with the product-market managers. Under functional authority, the product managers have the opportunity to communicate the importance of local adaptation via cheap talk; the functional manager determines the conditional expected value of implementation, and decides whether to impose or not synergies based on his incentives. The analysis of an organization with functional authority provides a rich description of the tradeo between incentives and coordination. Tightly linking pay to own unit performance strengthens incentives and increases e ort, while reducing incentives by loosening up that link improves communication and implementation. To understand this tradeo suppose that a watchmaker s earnings are tightly linked to his own performance. Then the watchmaker works hard, but he is at odds with the movement manufacturer he does not want to ever implement a standardizing action. Similarly, if the movement manufacturer s performance is tightly linked to his own performance, he never wants to pass up on synergies that are incongruent with local demands. As a result of this con ict of interest, communication (which is always cheap talk) cannot be credible. Moreover, implementation is suboptimal, as a manager is never willing to value equally his own and someone else s losses. If the link between pay and performance is weakened, i.e. the power of incentives is reduced, managers incentives are more aligned with one another, improving communication and implementation. Thus the tradeo between incentives and coordination takes two forms. First, as incentives become stronger, implementation by whichever manager has authority becomes more biased. Second, as incentives become stronger, credible communication is less possible, up to the 4

5 point where managers do not believe any of the information they are transmitted by managers with very strong incentives. This reasoning underlies our rst set of results. With functional authority, attaining synergies inevitably involves organizational costs in the form of (1) lost local adaptation and (2) weaker incentives. Moreover, if communication is desired, so that decisions about synergies take into account the relation between the value of synergies and the value of local adaptation, then incentives may need to be further reduced. We compare the value of functional authority with a decentralized organization in which the individual product managers always decide to adapt their products and synergies are ignored. We show that functional authority is most likely to be preferred when synergies are high; when the information by the local managers is valuable, that is, when making an ex-post decision on whether to adapt to the local circumstances or to impose synergies is important; and nally, when e ort the importance of e ort incentives is not too high As the importance of e ort incentives grows, the organization moves towards less ex-post decision-making; if the value of local adaptation is high or the value of synergies is low, this move takes the form of a move towards a decentralized organization. In section 5, we consider an organizational structure with product authority that may allow communication. We allow product managers to cede control to the functional manager after learning their private information. If both product managers do so, the functional manager can decide whether or not to implement the synergy. We show that there are two main cases. First, the organizational designer may only want product managers who face a low value of local adaptation to be willing to accept synergies. This organizational form may be optimal because it is inexpensive, in terms of reduced incentives, to induce this limited communication and the corresponding implementation, relative to communication and implementation under functional authority. Second, the organizational designer may want to induce some managers for whom the value of local adaptation is high to be willing to yield control. This case is very costly in terms of e ort incentives, as it requires su ciently balanced incentives that even a manager for whom local adaptation is very valuable prefers yielding control to simply refusing to implement synergies. We show that, in fact, this case is always dominated by functional authority. The consequence of this reasoning is that, as we show, product authority is an attractive organizational structure only when incentives are important and synergy implementation is only important with low values of local adaptation. One practical implication of our analysis is what it tells us about the relation between potential synergies and those that can actually be achieved. For example, in a merger between two companies, one should not rely on the headline value of potential synergies but must consider rst, the costs from reduced local adaptation, and second, the costs from reduced incentives to achieve second-best implementation decisions. Part of this calculation is that some potential synergies will not be implemented because the organizational costs are too high. Our model can provide insights into the size of this organizational discount. Consider for example the impact of the size of the synergies on the percentage that will be 5

6 realized. It is clear that there is a direct e ect the higher the synergy, the more valuable the merger. There is another e ect, operating through the organizational cost of implementing the synergies. The higher the synergies, the lower the organizational discount that must be applied to a merger, all else constant. The reason is that, as synergies get su ciently high, contingent decision-making is less important and so are balanced incentives. As the synergies are su ciently high, high powered incentives to the functional managers have few costs in terms either of bad implementation (the synergies must be most likely implemented anyway) and in terms of non-credible communication from product managers (it can be ignored with a high likelihood anyway). 1.3 Literature review Our analysis contributes to several theoretical literatures. First, it contributes to the literature on multitask incentives (Holmstrom and Milgrom (1991, 1994), Holmstrom (1999)). This literature models the tradeo among multiple tasks, some of which cannot be a ected by incentives directly, in a reduced form setting. Roberts (2004) argues, without a model, that the key multitask tradeo is the one between coordination and incentives (or initiative). Essentially, the problem is that organizational designs that are able to provide strong incentives (such as stand-alone organizations) are unable to coordinate the e orts of di erent agents, and vice versa, organizational designs that emphasize coordination (such as functional organizations) fail at providing incentives. This idea seems important, but no model we are aware of is able to rationalize and explore the power of this tradeo. In this paper, we both provide content to the multitasking intuition, and thereby deepen our understanding of how this tradeo operates. Second, our paper contributes to the literature on hierarchical information aggregation. In that literature, managers can integrate the information of subordinates. The reason a manager is valuable is because he can make decisions which are based on more information than each individual agent. 5 This literature assumes that there is no strategic component to the information transmission of the managers subordinates. In this paper, we consider the realistic situation in which, for the manager to be able to integrate more information, it must be in the interest of the other agents to transmit that information. In this sense, we follow Dessein (2002,2003) in modelling communication in organizations as a cheap talk game. The allocation of control then a ects both the nature of communication and the use of the resulting information. Other models which explicitly analyze the impact of control on information aggregation are Aghion and Tirole (1997) and Harris and Raviv (2003). The focus of the above papers, however, is on delegation of authority as a way to make better use of dispersed information. In contrast, in our model, e ective aggregation of information often requires centralizing control. Moreover, we endogenize the incentive con icts which are at the origin of this strategic communication. 5 See e.g. Keren and Levhari (1983), Radner, (1993), Van Zandt (1999), Bolton and Dewatripont, (1994). 6

7 Third, our work contributes to the literature on how organizational structure tries to capture synergies. Scholars have addressed organizational design as a way to solve coordination problems at least since Chandler s (1962) classic analysis of the impact of the change from functional to multidivisional organizational forms on the performance of large corporations. Also seminal is Lawrence and Lorsch s (1967) study of the mechanisms organizations use to integrate its di erentiated activities. The theoretical literature trying to understand these coordination problems originated with the theory of teams of Marshack and Radner (1972). This approach studies coordination among agents when information must be necessarily disperse, but purposely assumes away incentive issues to simplify the analysis. 6 Agents with common interests would choose to share their information if they could, so an additional necessary restriction is the existence of some communication costs that prevent agents from freely communicating. A few papers within this approach have taken into account the existence of intendependencies among units. Cremer (1980) studies the optimal grouping of subunits into units in a resource allocation problem. The existence of interdependencies among subunits means that any particular grouping creates externalities, as some interdependence must be ignored. The question he considers is how to group units in a way that internalizes as many of these externalities as possible, when agents ability to collect information is limited. The attractive generality of the approach limits its insights to the initial one that the organization should put together related units and segregate unrelated ones. 7 Harris and Raviv (2002) study the organizational structure that best appropriates synergies when managers are expensive. Their model, however, abstracts from con icts between synergies and from the use of local knowledge for local adaptation. Most closely related to our model, Roland, Qian and Xu (1999) model the tradeo involved in the organizational design choice as one between decentralizing and allowing multiple divisions to use their local knowledge versus losing standardization and economies of scale. Among other things, they show that the M-form improves coordination at the expense of economics of scale. While our approach coincides with theirs in modelling a con ict between realizing synergies and adapting to local knowledge, they do not consider the tradeo between incentives and coordination. Moreover, communication failures are exogenously imposed, and not the result of an endogenous incentive con ict as in our model. Fourth, our paper contributes to a small, recent literature that has studied jointly the incentive problems and the coordination costs that follow from the design decision, focusing in 6 Team theory models assume that each player has the same preferences with respect to outcomes and actions. The papers described in the paragraph on hierarchical information aggregation also fall within that team theoretic tradition, but assume away interdependencies. 7 Aoki (1986) and Genakoplos and Milgrom (1991) deepen and generalize the study of a resource allocation problem of this kind. Vayanos (2002) studies decentralized information processing when tasks are related. In his setting, when agents aggregate information to communicate it to a superior, some information is lost. Moreover, the decision of one agent impacts those of other agents. Under these circumstances, the problem is to determine how best to group units into subunits to minimize the information loss while taking into account the interdependencies. 7

8 particular on the M-form versus U-form choice. Holmstrom and Tirole (1991), analyze transfer pricing between di erent divisions under interdependencies. They associate the M-form with a decentralized organization in which division managers are free to both trade internally and with the external market. They show that, while the M-form tends to maximize incentives, it results in divisions being less well coordinated relative to more centralized organizational forms which do not allow external trade. The problem they study is a pure moral hazard problem; the informational consequences of the design play no role in it. Maskin, Qian and Xu (2001), in contrast, highlight the advantages of the M-form in providing incentives based on yardstick competition, but interdependencies between decisions play no role. Hart and Moore (2004) study how to allocate authority over the use of assets when agents with several assets (coordinators) can have ideas involving the common use of several of these assets, and when agents are motivated by their own interest rather than the organizations. Our work shares with theirs the focus on the allocation of authority, but we are concerned, unlike in their work, with communication agents can communicate their preferred choices, if it is in their interests, and that information may be aggregated. We also study the incentive con ict explicitly as a consequence of the incentives for e ort provision that must be provided to the agents that is the parties can reduce this incentive con ict through the design of alternative incentive mechanisms. Also related, Hart and Holmstorm (2002), in a framework centering on the managers private bene ts of control, argue that whereas independent rms coordinate too little their activities, integrated rms have a tendency to realize too many synergies, neglecting private bene ts of mangers and workers. While this paper shares our view that organizational structure a ects incentives, in it, unlike in ours, organizational form does not a ect the use and availability of information. Finally, and closest to our work, Athey and Roberts (2001) focuses on how the allocation of authority a ects the tradeo between giving agents incentives for decisions and for e ort provision in a situation where, while those components are additively separable in production, only a broad signal that adds both incentives and the output from the project is available. We di er from their approach in two main aspects. First, our emphasis on coordination there are no synergies in their case, as projects do not interact with one another, unlike in our paper. Second, communication between agents of the information agents obtain is impossible in their analysis a boss can learn the project payo s at an additional cost but with no distortion. Communication, as determined endogenously by the available incentives, plays a key role here. Thus also the aggregation of the information of the di erent agents, is absent from their analysis but plays a central role in organizational design in ours. More broadly, what distinguishes our approach from previous papers is the ability of our model to study organizational design issues when agents are self-interested and coordination among them is important. In our view, developing a framework that can deal both with the reasons the organization is actually set up as well as with the informational asymmetries and 8

9 incentive con icts that emerge as a result of the design decision is an important step towards a deeper understanding of both organizations and incentives. 2 The model Production Coordination: Consider two agents each specialized in supplying a product. They recognize that there are gains from combining a functional activity such as production, research and development, marketing, or distribution. For example, there may be cost-savings in manufacturing from producing the same product in a single plant. Implementing synergies requires a third person, the functional manager, who specializes in identifying the speci c opportunity and coordinating the joint functional activity. We model this by assuming that the functional manager learns the value of an independent random variable k v U[0; K]; that determines the payo from coordinating the activities of both specialized agents. Adaptation: The cost of implementing the synergy is that standardization results in an individually suboptimal choice for the each product and each product manager may need to alter its decisions to accommodate the functional manager s choice. For example, if the functional manager oversees production and adjusts product design to increase standardization, it may result in a product that is suboptimal from a product manager s perspective, and the product manager may need to alter its marketing or distribution plan to match the new product design. How much product market pro ts decrease is given by the costs of adaptation,. One can think of as representing the cost of not being adapted to the local environment, that is of producing a product that is not ideal for local market conditions. is measured by an independent and identically distributed random variable drawn by each of the two product divisions, which we denote 2 f l ; h g; > 0: Producing a non-adapted product costs ; while producing a product that is adapted yields 0: We let p be the probability that local conditions are important so the value of adaptation is high, that is: Probability[ = h ] = p. We assume further for simplicity (to reduce the need to track di erent cases from corner solutions throughout the analysis), that K > 2 h, the maximum available product synergies are higher than the maximum value of adaptation. E ort: Agents must also exert unobservable e ort, which entails a private cost. Thus, the extent to which a manager keeps the return from his unit a ects his e ort choice. For simplicity, we assume returns from e ort are separable from coordination and adaptation returns. In particular, the marginal product of e ort is given by v j e j ;where v j is a unit-speci c parameter for functional or product unit j. Agents information and incentives Managers obtain private information about the units they oversee; product managers about adaptation bene ts ; and functional managers about coordination bene ts k. If no functional manager is involved, no information about synergies is obtained. We assume that agents are risk neutral, so that absent communication 9

10 and implementation issues, the incentive problem can be simply solved by allowing agents to keep the entire product of their e ort. Calling w the agent s earnings, the utility of the agents then takes the form: U = w e2 2 : Contractibility and observability We assume that e ort of each agent is unobservable, but that each unit s pro ts (total costs for functional unit) are observable. The pro t on each product is given by the e ort of the agent working on the product minus the potential loss of adaptation: R i = R + v i e i i I; for i = 1; 2; where I is an indicator variable that describes whether the synergy is implemented. Total functional costs depend on whether or nor the synergy is implemented and the functional manager s e ort of the coordinator, are given by: C = C ki v f e f : R and C are constants that we can normalize to 0; since both ex ante and ex post uncontingent transfers are allowed, the constants play no role in the analysis. Total organizational pro ts then are: f = X e 2 i (v i e i 2 i=1;2 i I) + ki + v f e f We assume that the synergy implementation decision is unveri able. e 2 f 2 Since only total product unit pro ts and total functional costs are observable, contracts can only be written on these variables. We further restrict contracts to linear shares of these variables. Linear shares are unlikely to be optimal given that the synergy implementation is a 0-1 decision and the adaptation costs are binomial, it should be possible to make inferences about e ort from the distribution of contractible outcomes. Such inferences will be impractical if there is additional noise in the observables; since we assume risk-neutrality, the noise would a ect nothing else in the model. Let s f represent the share of functional costs that the functional managers receives and s i represents the share of product pro ts that a product manager receives. We assume that the contracts for the two symmetric product managers are the same and budgets must balance in every state of nature, so s 1 = s 2 = s p and the functional manager gets a share 1 a share (1 s p of each product unit s pro ts and each product manager bears s f )=2 of the functional unit s costs. We do not allow one product manager to obtain a share of the other product unit s pro ts. Thus, the functional manager seeks to maximize 10

11 f = s f (ki + v f e f ) + (1 s p )( 1 I 2 I + v p e 1 + v p e 2 ) and each product manager seeks to maximize e 2 f 2 i = s p ( i I + v p e i ) + 1=2(1 s f )(ki + v f e f ) e 2 i 2 : Organizational structure: communication and control In designing the organization of these three agents, the organizational designer faces three problems: providing the right incentives for e ort, facilitating credible communication between the agents, and ensuring that whichever agent makes the synergy implementation decision has incentives to decide in the way most favorable to the whole organization. Of course, as we will show, these three objectives are in con ict, and the organizational design must manage the tradeo s optimally. The tools available to solve these problems are the design of the incentives provided, which we summarize by the shares of the output they receive; the allocation of managers to functions; and the allocation of decision rights to these managers. An organizational structure in our model is an allocation of decision rights over synergy implementation, a communication protocol, and a sharing rule. All communication is cheap talk; it is unveri able so no contracts can be written that depend on messages. For each combination of decision rights and communication protocol, we solve for the optimal incentives. We analyze three distinct structures. rst, we study ex-ante decision-making, where the organization allocates decision rights to either the product managers or the functional manager and there is no communication. Second, we analyze functional authority where the product managers can communicate their information to the functional manager who in turn decides whether or not to implement the synergy. Third we study product authority where the functional manager communicates whether or not he wishes to have the synergy implemented and the synergy is implemented if both product managers want implementation. 3 Ex-ante decision-making The organization may opt for ex-ante decision-making, by imposing a simple ex-ante decision rule. This may be of two types: a decentralized structure that ignores the synergy for any information realization or a structure that implements the synergy for any information realization. In either case, all three managers will optimally be given rst-best incentives. The absence of incentive con icts re ects the fact that in this case it is easy to recognize which manager s contribution is responsible for which output. If the organization never implements synergies, product managers always adapt their products to the local environment, so each manager s problem is given by: i D = max v i e e e

12 which results in rst best e ort e = 1;with output v i =2: The organization is perfectly adapted and it does not incur any non-adaptation costs. Total organizational pro ts are then D = v 2 p + v2 f 2 : Suppose instead the organization always implements synergies. Again, each manager is given rst-best e ort incentives and chooses e = 1: The expected value of functional synergies is K=2; and if we let be the expected costs of not adapting for each product, then pro ts in this case are: XS = K v 2 p + v2 f 2 : The di erence in pro ts between both forms is then simply: D XS = 2 K 2 : That is, the preferred form simply depends on whether or not the expected value of synergies exceeds the costs of not adapting to local conditions. Both of these organizational forms implement an ex-ante decision rule. First-best implementation demands ex-post decision rules. All the information should be used to determine whether or not to implement synergies, but this requires communication. We consider next how this may be achieved. Before doing this, we further simplify the model. Our analysis is concerned mainly with the importance of the synergies versus local adaptation, and for this reason we will assume throughout, that e ort is equally productive in both activities, that is v 2 f = 2v2 p = v: That means that, for equal e ort inputs, total e ort output produced by the two row managers equals the e ort output of the column manager. This implies that, reducing either s p or s f by a given amount then has the same impact on total e ort output. Assumption 1: v 2 f = 2v2 p = v: 4 Functional authority In this section, we analyze the organizational structure in which the functional manager has the authority to decide whether or not to implement synergies and the product managers have the opportunity to communicate their information to the decision maker. The game proceeds as follows: the product managers simultaneously have the opportunity to send a message to the functional manager that, depending on the parameters and incentives, may credibly reveal their private information. The functional manager then decides whether or not to implement the synergy. Since the returns from e ort are separable from the implementation decision and the private information and shares are linear, the timing of the e ort decision does not matter. 12

13 A priori, it appears this organizational form may allow the use of ex-post information on the relative value of functional synergies versus product adaptation, and as a result may increase pro ts. However, suppose that we have the same incentives as above, that is the product managers enjoy the entire value generated by the local adaptation and the functional manager receives the functional payo s. Then it is clear that no useful communication can exist. Moreover, the functional manager always will want to implement the synergy; the synergy is always non-negative and the functional manager bears none of the costs from lack of adaptation. This result motivates the analysis that follows, and thus we present it formally in the following proposition. All the proofs are in the appendix. Proposition 1 If rst-best e ort incentives are provided to all managers (s f = s p = 1), no communication takes place. Moreover, the decision of the functional manager is to always implement the synergy, and no ex-post information is taken into account. Simply altering the organizational design by allowing communication has essentially no impact if the incentives remain what they were in the ex-ante organization. The reason is that both credible communication and implementation decisions sensitive to the size of synergies and bene ts of adaptation require that the aims of the managers involved not be completely opposed. Strong incentives are thus an obstacle to the ability of the organization to implement tradeo s between synergies and adaptation. Reducing the share of output that each manager obtains from his or her own unit may mitigate this problem. First, conditional on the communication that takes place, such a reduction in e ort incentives improves implementation, by giving the functional manager a stake in the organization-wide bene ts that obtain when there is better local adaptation. A functional manager that pro ts somewhat from the product-adaptation decision may sometimes decide to give up on implementing the functional synergies (when k is low) and agree to allow local adaptation by the product managers. This would be the case even if no truthful communication from the product manager occurs. Second, if incentives are su ciently aligned, communication where product managers can credibly represent the costs to them not adapting to local conditions, may be possible. In the next section we consider the rst of these problems: ignoring the possibility that truthful communication can take place, design incentives so that implementation is improved, at the cost of reducing e ort incentives. 4.1 Ex-post decision-making without communication: implementation versus incentives Consider implementation decisions when the only information the functional manager has is the level of synergy k and the ex-ante expected value of adaptation. Intuitively, it is clear that there are shares that support some degree of ex-post implementation restraint, that is 13

14 if the realization of the value of the synergy, k; is low, then the manager will choose not to implement the synergy and will instead allow for local adaptation. Since the functional manager has no information on the cost of local adaptation he only compares the expected cost to him of losing local adaptation against the value to him ex post of the synergies realization k: That is, the optimal policy function for the organization consists of a cuto k pool such that if k < k pool ; the functional manager does not impose the synergies and the product managers are able to adapt locally, while if the synergies are high enough, k > k pool ; he implements the functional synergies. k pool is the value of the synergies at which the manager is indi erent between implementing and not implementing the synergies given the expected value of the adaptation by the product. Since the functional manager obtains a share s f of his own synergies, and su ers a share (1 s p ) of the cost of lost local adaptation for each product, k pool solves: s f k pool 2(1 s p ) = 0; which implies: k pool = 2(1 s p) s f (1) E ort decisions by the product and functional managers, respectively, are determined by: max s p v p e e p e 2 2 ; max e f s f v f e e 2 f 2 ; so e p = s p v p ; e f = s f v f : Now consider the expected pro ts of the organization that allow us to determine the optimal choice of s f and s p : The output attained has, as before, two components: the expected value of implementing the synergies, which is the value of the synergies minus the cost of local adaptation; and the value of the e ort incentives. That is, the optimization problem of the organization is (recall that v 2 f = 2v2 p = v): F (1 p) 2 p = max s f ;s p K + p2 K Z K Z K k pool (k 2 L )dk + 2 (1 p) p K Z K k pool (k 2 H )dk + vs p (2 s p ) + vs f (2 s f ) k pool (k L H )dk + (2) Substituting the value for k pool into the optimization, taking rst order conditions and sim- 14

15 plifying, the following expressions determine the optimal incentive choice: F ps p (s p; s f ) = v(1 s p) + (1 s p s f ) s 2 f K (2) 2 = 0 (3) F ps f (s p; s f ) = v(1 s f ) + (1 s p s f ) s 2 f K (1 s p ) s f (2) 2 = 0 (4) From these rst-order conditions, two immediate results obtain. First, the incentives of agents are not rst-best, i.e. improving implementation requires distorting e ort incentives. Second, perfect balance is not achieved, that is, one and possibly both agents receive more than half of their own output. Thus the organization trades o better implementation of the synergies decisions against better incentives for the agents. Lemma 1 The agents outputs shares are such that s p < 1 and s f < 1; so that both agents produce less than rst-best e ort. Moreover, incentives are never perfectly balanced, that is s p + s f > 1: Intuitively, at s i = 1 we have rst-best incentives for e ort, and a decrease in s results in a second order loss in e ort incentives and a rst-order gain in implementation incentives (because the lower bound of the distribution of k is 0); while at s i = 1=2 we have rst-best implementation incentives, and an increase in the shares results in a rst-order gain in e ort incentives and a second-order loss in implementation incentives. Thus in order to ensure some ex-post decision-making, the organization may choose to distort somewhat e ort incentives even if communication is impossible. The distortion s only aim is to improve implementation, by providing each agent a stake in each other s output. The next proposition characterizes the determinants of the extent of this distortion, and the role that synergies play in it. Proposition 2 (Incentive loss to improve implementation) Managers own e ort incentives are stronger, (i.e. s p and s f are higher) the higher the importance of non-contractible e ort (v), the higher the expected value of synergies (K) and the lower the mean cost of adaptation (). That is, ensuring good ex-post implementation by the functional manager is less relevant whenever the synergies are high, when the value of e ort incentives is high, or whenever the cost of adaptation is low. Note that a mean-preserving spread in has no impact on incentives, since, given that there is no ex-post communication, the actual value of the adaptation parameter (whether l or h ) is irrelevant to the functional manager s decision. 4.2 Incentives for communication Decision-making may be improved if incentives are chosen not just to ensure that implementation is optimal given the expected value of adaptation, but also to ensure that the decision 15

16 incorporates information about the actual realized value of adaptation. This requires that the incentives be su ciently aligned that the product managers can credibly communicate their observations on the value of adaptation for their unit. The question to explore once we allow for communication is to what extent we move towards more balanced or more high-powered incentives. Intuitively, the organization cares more about better implementation decision once the managers are more informed. The following section explores this intuition. The organization design problem is to encourage communication by aligning interests without incurring too severe e ort costs. Communication can be improved by giving product managers a stake in the functional synergies, and giving the functional manager a stake in local adaptation so implementation decisions will take into account product division costs. We proceed in two steps. First, we explore the incentive compatibility constraint that determines whether communication takes place for a given set of incentives. Then, we set up the optimization problem given these constraints, and characterize the solution. Incentive compatibility Suppose that product managers can send credible messages about the value of adaptation for their respective products, that is, suppose that a separating equilibrium exists. Then, similarly to the case when no communication is possible, the functional manager implements synergies if the payo he receives from the synergies is higher than the cost he incurs of not adapting the product to the individual local market conditions for each product. That is, synergies are implemented if: s f k (1 s p )( ) > 0: The realized values i are now known since we assume separation. Thus this condition determines a decision rule with three cuto points: k LL ; k LH and k HH ;which equal the level of k for each level of s, such that the functional manager implements the synergies if cuto value. The cuto s then are given analogously to the previous section by: k is above the k ij = (1 s p)( i + j ) s f (5) with i; j = L; H: The rst best implementation rule is k fb ij = ( i + j ): Thus the extent to which we have ine ciently too much implementation or too little implementation by the functional manager depends on whether 1 s p? 1; i.e. s p + s f? 1:In particular, if s f 1 s p < 1;(as we show below is indeed the case) we will have that k ij < k fb ij ; and the s f functional manager implements synergies too often and allows for too little local adaptation. We can now write the truth-telling constraint of a product manager who must decide whether or not to truthfully report the cost of implementing a synergy in terms of lost local adaptation. When the product manager sends a message to the functional manager, he knows nothing about the value of synergies. To decide whether truthtelling is in his interest, he must form an expectation over k: To write the incentive compatibility constraint, note that the product manager who is tempted to lie is one for whom adaptation is not valuable, since it is 16

17 in that case that the synergies will be more likely to be implemented by the column manager. Truthfully reporting L is preferred if: (1 p) K (1 p) K KZ k LL KZ k LH (1 s f ) k 2 (1 s f ) k 2 s p L dk + p K s p L dk + p K KZ k LH KZ k HH (1 s f ) k 2 (1 s f ) k 2 s p L dk s p L dk: The left-hand side of this inequality is the payo of correctly communicating L. In this case, with probability (1 p) the other agent also reports L ; in which case the probability of synergies being implemented is the probability that k > k LL ; while with probability p the other agent reports H ; in which case the probability of synergies being implemented is the probability that k > k LH. On the other hand, if the agent lies, the integral is taken over a smaller set of k s: in the rst case (if, the other agent draws L ) for values k > k LH ;in the second (when, with probability p; the other agent draws h ) over k > k HH : That is, the value of lying is in the increase in the value of k that the functional manager has to observe before he decides to implement synergies. The integrals simplify in the obvious way, and the IC constraint becomes: (1 p) K kz LH k LL (1 s f ) k 2 s p L dk + p kz HH (1 s f ) k K 2 k LH s p L dk > 0; (6) where the value of the cuto k ij is given by (5). We further simplify this constraint under the assumption that p = 1=2. In this case, and after some simple manipulation, the IC constraint becomes: (1 s f )(1 s p ) s p s f 2 L L + H (7) Thus, for given shares, the IC constraint is more likely to bind when H and L are close to each other. Second, for a given pair of s; the constraint is less likely to bind when the left hand side is higher, that is when the shares are more balanced. In particular, if both shares are 1=2;the left hand side of the IC constraint is 1;and truthful communication is always incentive compatible regardless of the values of : We next consider the optimal choice of incentives in two cases. First, when the optimal choice of s p and s f for implementation is such that the constraint is not binding in this case communication does not involve further distortions beyond the implementation distortions. Second, we analyze the case where communication requires that the constraint binds. The optimization problem The organizational design problem involves choosing incentives that optimally trade o implementation and e ort conditional on credible communica- 17

18 tion being supported. The expected pro ts of the organization are given by the probability that synergies are implemented times the expected net output of synergies plus the value of the e ort, which is a function of the shares. The program is to maximize: F sep = (1 p)2 K + p2 K KZ KZ k LL (k 2 L ) dk + 2 (1 p) p K k HH (k 2 H )dk + s p (2 s p )v + s f (2 s f )v KZ k LH (k L H ) dk (8) subject to (7). As before, the marginal value of increasing the output shares is given by the incentive value of the share the extra e ort agents provide. The marginal cost of higher incentives results from the increasing con ict in the implementation decisions that leads to worse implementation the less balanced the shares, the more implementation decisions depart from simply comparing the value of the synergies with the value of local adaptation. Whether the constraint (7) binds is crucial for how incentives will be chosen. If the constraint does not bind, the rst-order conditions are: F seps p (s p; s f ) = v(1 s p) + (1 s p s f ) s 2 f K E( ) 2 = 0 (9) F seps f (s p; s f ) = v(1 s f ) + (1 s p s f ) s 2 f K (1 s p ) s f E( ) 2 = 0 (10) Essentially the same forces are at play as in the pooling case, except that where only the ex-ante mean of the cost of adaptation matters in the pooling case, now the variance in is important, as the decisions on synergy implementation are taken ex post. However, if (7) binds, then the two shares move in opposite direction to one another; it is easy to see this by simply solving for s f in (7). Intuitively, ensuring communication when the constraint binds means that a higher share of his own output for a product manager will have to be counterbalanced with a higher share of the functional product division output in order to preserve the balance in incentives that the product manager needs to communicate truthfully. The following proposition characterizes formally the comparative statics in these two cases. Proposition 3 (Synergies and Incentives with Communication) (1) In a separating equilibrium where the IC (6) is non-binding, s p and s f are increasing in v and K and decreasing 18