تخصیص مجدد منابع شرکت ها و مشوق های مدیریتی در بازارهای سرمایه داخلی

Diversified firms often trade at a discount with respect to their focused counterparts. The literature has tried to explain the apparent misallocation of resources with lobbying activities or power struggles. We show that diversification can destroy value even when resources are efficiently allocated ex post. When managers derive utility from the funds under their purview, moving funds across divisions may diminish their incentives. The ex ante reduction in managerial incentives can more than offset the increase in firm value due to the ex post efficient reallocation of funds. This effect is robust to the introduction of monetary incentives. Moreover we show that asymmetries in size and growth prospects increase the diversification discount.

The analysis of the allocation of funds among different divisions of a conglomerate firm is a relatively young topic. Stein (2002b) provides a recent survey. The general theme coming from the empirical literature is that diversified firms trade on average at a discount relative to a portfolio of focused firms in the same industries, as reported by Berger and Ofek (1995), Servaes (1996) and Lins and Servaes (1999). Moreover, the 1980s saw a process of dismantling of diversified firms, driven by the idea that the divisions would be more efficiently managed as stand-alones. But if there is by now a wide consensus on the idea that a diversification discount exists, it is much less clear why this is the case. Stein (1997) has pointed out that internal capital markets can create value in financially constrained firms. In Stein's words, “Simply put, individual projects must compete for scarce funds, and headquarters’ job is to pick winners and losers in this competition.” Stein denotes this activity of headquarters in a conglomerate firm as “winner-picking”. Contrary to the empirical findings, Stein's model suggests that internal capital market should create value and thus a premium for diversified firms. One possible way to solve this apparent paradox is to argue that the discount of diversified firms is due to misallocation of resources in internal capital markets. For instance, Rajan et al. (2000) find that multi-segment firms allocate relatively more capital to “weak” lines of business than their stand-alone counterparts, and relatively less to segments in “strong” lines of business. Scharfstein (1998) finds that the investment of conglomerate divisions is virtually insensitive to investment opportunities, as measured by the industry q's. Lamont (1997) shows that resource allocation in diversified firms is different from that in focused firms and less sensitive to indicators of investment value such as Tobin's q. However, the evidence on inefficient allocation of funds has been disputed. Whited (2001) points out that the inefficiency results appeared in the literature may actually be due to the incorrect measures adopted for the investment opportunities of the divisions. She shows that when measurement problems are taken into account, the evidence of inefficient allocation of funds disappears. Chevalier (2000) analyzes the investment behavior of a sample of firms before and after diversifying mergers, finding no evidence of a change in investment behavior. This implies that, if there is inefficiency in the investment behavior of the divisions of conglomerate firms, such inefficiency does not appear to be due to the presence of internal capital markets. In this paper we argue that in order to explain the diversification discount we do not need to assume any misallocation of funds in internal capital markets. Conglomerates can destroy value even if resources are efficiently allocated. If managers derive utility from the funds under their purview, the possibility of implementing a “winner-picking” policy, while optimizing resources allocation ex post (i.e. after managerial effort has been exerted), reduces managerial incentives to exert effort. Taking away from the manager the cash flow generated by the division has the negative implication of reducing the incentives for division managers to spend effort to generate the cash flow. The reduced managerial incentives can more than offset the gains of reallocating funds to the most profitable divisions. In other words, “winner-picking” is both the bright and the dark side of internal capital markets. We consider a two-period model with two divisions and a headquarters. Division managers receive private benefits in proportion to the gross return of the division they run. Headquarters maximizes total firm value. In the first period the two division managers have to exert a non-verifiable effort to increase the probability of success of a project already in place. The cash flow generated by the existing project will be reinvested inside the firm in the second period. Before the second period, the headquarters receives a signal on the second period profitability of the two divisions and reallocates funds. When divisions operate as stand-alones, each division reinvests the cash flow generated by the first period project. On the contrary, in the diversified firm the headquarters will redistribute the cash flow to the most profitable division. The redistribution has two effects: on one hand it creates value, but on the other hand it reduces the rent for the manager of the (ex post) less profitable division. Anticipating the possibility of being expropriated, each division manager will reduce his effort. Consequently, the total cash flow to be reinvested in period two will decrease. This observation has the following implications. First, a profit-maximizing headquarters will face a time inconsistency problem. Once the funds are generated, headquarters would like to exercise “winner-picking” to the highest extent possible. However, this ex post maximizing behavior by headquarters will reduce ex ante incentives at the divisional level, and it may cause a loss of value for the corporation. When the gain from reallocating funds across divisions is limited, i.e. when “winner-picking” has a limited potential for creating value, the negative effect on the reduced managerial incentives dominates and the diversified firm trades at a discount. Conversely, when the gains obtained from reallocation of funds across divisions are large, then the advantages of “winner-picking” dominate over the reduced managerial incentives, and the diversified firm trades at a premium.1 Second, diversity in the ex ante profitability of divisions increases the likelihood that a conglomerate trades at a discount. If one division has a very high probability of having ex post the best investment opportunities, the incentives for the manager of the other division are reduced. Therefore the cash flow that can be reallocated to the most profitable division is also reduced, limiting the gains of “winner-picking”. Finally, the diversification discount is greater when a division with a greater potential for immediate cash generation is paired with a division with poor capacity of immediate cash generation but good growth prospects. In this case the manager of the first firm will fear expropriation of the cash flow generated in favor of the high-growth firm, thus reducing the conglomerate's value. The basic intuition that ex post interference by the principal may be harmful for the agent's ex ante incentives is of course not new. For example, Aghion and Tirole (1997) show that if the principal intervenes too often in the decisional process, it can stifle the agent's initiative. Gertner et al. (1994) point out that giving control rights to capital providers in an internal capital market may be costly in that it diminishes managerial incentives. The manager of a division is more vulnerable to the opportunistic behavior by corporate headquarters than a manager of a firm receiving financing either from a bank or from an external financial market because headquarters have control rights over the division's assets. Contrary to a bank, headquarters can liquidate the assets even when the division performs well. In their model the hold up problem between headquarters and divisions holds irrespective of the possibility of reallocating resources across divisions. In our model it is precisely the “winner-picking” ability of the headquarters that blunts managerial incentives.2Rotemberg and Saloner (1994) discuss how, in the presence of incomplete contracts, firms may benefit from restricting the scope of their activities. Their basic idea is that diversified firms have a wider range of projects to implement, and for some reason they cannot implement all of them. As a result, they are more likely to implement a project that it is not ideal for providing ex ante incentives. They also propose an application of their model to internal capital markets, showing that it may be optimal to force each division to use only funds that it has generated itself. We extend their argument pointing out the cases where internal capital markets are less likely to be beneficial. Other papers which present arguments similar to ours include Gautier and Heider (2000), Inderst and Laux (2000), Inderst and Müller (2003) and De Motta (2003). Gautier and Heider (2000) analyzes a model in which division managers must first produce cash flow that will be then reinvested inside the company. However, in their model projects have a fixed size (an extreme version of decreasing marginal returns of capital) and consequently the scope for winner-picking is reduced. In Inderst and Laux (2000) effort is directed to the generation of investment opportunities rather than cash, and their focus is on the role played by liquidity constraints. Inderst and Müller (2003) is also focused on the impact of conglomerates on financing constraints. De Motta (2003) studies a model in which division managers try to influence the internal capital market's assessment of their division in order to boost their level of funding. In his model the difference between external and internal capital markets is the informational advantage of the latter, while in our model the distinctive feature is the headquarters’ ability to reallocate funds across divisions and informational asymmetries do not play any role. Mailath et al. (2002) apply a similar setting to study the cost of mergers. In their model there may be disadvantages in merging two firms even in the case when the merger allows the internalization of externalities between two firms. The reason is that the redeployment of assets implied by the merger can increase the cost of inducing managerial effort by making unprofitable certain decisions like the termination of an unprofitable project which has a positive externality on the other divisions. Hart and Holmström (2002) consider a firm with two divisions and compare two different organizational model. In one case decisions are taken at the divisions’ level. The disadvantage of this model is that the division's manager does not take into account the externalities created to the other division. On the other hand, the advantage of the decentralized decision making is that the private benefits of the division are considered. When decision making is centralized the opposite occurs: total profits are considered, but no weight is put on the division's private benefits. On the empirical side, a number of papers have tried to explain the “diversification discount” as the consequence of a non-random selection of firms that become conglomerate. Papers in this literature include Campa and Kedia (2002), Fluck and Lynch (1999) and Maksimovic and Phillips (2002). The general point is that weaker firms may have a comparative advantage in merging. Thus, even if conglomerate firms work efficiently, a diversification discount appears. Our point is different, and it is related to the different managerial incentives provided by conglomerate firms. As a final remark, we stress that we do not want to argue that resources are indeed allocated efficiently in internal capital markets. Power struggles, influence activities etc. are surely present in most corporations, and such inefficiencies contribute to reducing the value of diversification. Our point is that diversified firms may well trade at a discount even if internal capital markets allocate funds efficiently. The optimal policy ex post is not necessarily the optimal policy ex ante. The paper is organized as follows. In Section 2 we describe the basic model. Section 3 illustrates the main effects of the “winner-picking” policy, comparing the performance of a diversified corporation in which funds are allocated ex post efficiently with the performance of a “stand alone” firm. In Section 4 we show that the basic results still hold when the firm can provide monetary incentives to the managers. Section 5 discusses the impact of asymmetry in size between the two divisions. Section 6 contains the conclusions, and an appendix collects the proofs.

This paper has argued that one of the distinctive features of internal capital markets, that is the ability of headquarters to reallocate funds across divisions (winner-picking) is associated both with costs and benefits. The benefits derive from transferring funds to the most profitable divisions; the costs derive from the weakening of managerial incentives. In other words, winner-picking is simultaneously the dark and the bright side of internal capital markets. Our theory can explain why conglomerate firms trade at a discount (or at a premium) with respect to their focused counterparts. More importantly, it does so without assuming any inefficiency in the allocation of corporate resources. We show that ex ante diversity in divisions’ profitability increases the inefficiency of an internal capital market, and that mismatches between divisions’ size and profitability also reduce the value of internal capital markets. An important caveat is that we have not addressed the reasons why divisions that are very different in terms of their profitability are brought together in the same firm and why some divisions are not spun-off in those circumstances where conglomerates are inefficient. Moreover, we have assumed that all resources are internally generated, ignoring the role of external financing. These limitations notwithstanding, we believe that the analysis of internal capital markets in terms of allocation of delegation of authority may be a promising direction for future research.