انتخاب سازمانی و برونی در فرانچایزینگ

In this paper, we examine some implications of externality for the organization of firms. The need to control externality explains the selection, at the level of the chain, of full integration, dealerships or franchising systems, or systems of dual distribution where company and franchised outlets operate simultaneously, in preference to unrestricted retailing. We show that there could be a trade-off between managerial motivation and effective controlling of externality. This trade-off can explain the selection of particular organizational structures within franchising. In particular, non-separable externality, where the value of the externality depends upon characteristics of both the generating and affected unit, is costly to control contractually and could encourage integration.

In this paper, we examine some implications of externality for the organization of firms. In particular, we offer an explanation of franchising characteristics including the practice of dual distribution, where a franchise chain simultaneously uses both franchised and company-operated outlets. The need to control externality is cited as an explanation of the selection of full integration, dealerships or franchising systems in preference to unrestricted retailing (Rubin, 1978; Williamson, 1981 and Williamson, 1985). We show that the relative costs of inefficient effort and of controlling externality can also explain the selection of franchising in relation to dual distribution. In particular, non-separable externality, where the impact of the externality depends on characteristics of both the generating and affected satellite business, is costly to control contractually and makes it likely that integration will be selected. We begin by examining externality issues in franchising, also giving some attention to examples of externalities based on interacting effort levels. We also consider the interaction between the encouragement of localized business efficiency and the control of externality. Our formal model is based on the trade off between managerial motivation and externality control and leads to empirical predictions. 1.1. Externality issues in franchising Franchisors save on many day-to-day monitoring costs owing to the profit incentive for the franchisee to control local business costs (Maness, 1996). Nonetheless, the franchisor must monitor the effort of individual franchisees to prevent them from horizontally free riding on the reputations of other franchisees (Rubin, 1978). Others have seen franchise royalties as motivating the franchisor to control a vertical externality, where franchisees might try to free ride upon centralized promotional efforts (Mathewson & Winter, 1985). An externality problem potentially exists in any retail network and is likely to have organizational impact (Dnes, 1993; Williamson, 1981). The externality can in principle have positive or negative characteristics. Franchise systems encounter problems of motivating franchisees to honor guarantees and service arrangements issued elsewhere in the system (Dnes, 1992). System-wide customer-servicing arrangements can increase sales for the chain as a whole but will impose costs on franchisees. Such an externality can have negative (effort-reducing) effects as franchisees try to free ride, and positive (demand-enhancing) effects. Franchisees try to avoid responsibilities that they believe will just cost them money, ignoring the system-wide effects of this behavior. As it is in the interest of the system as a whole to have transferable service arrangements, there are classic free-rider aspects to this problem.1 The problem of franchise externality is particularly severe where the impact of an externality depends upon some aspect of the satellite business it affects, i.e. where it is non-separable. An earlier literature shows that non-separable externality, where marginal external cost cannot be defined independently of the level of activity of the ‘victim’, is much more difficult to internalize than the separable kind (Davis & Whinston, 1962). In the non-separable case, the cross-partial derivatives of the output function are not zero with respect to the externality and other inputs. Broadly, non-separability implies that some form of integration is much more likely, whereas separable externality may be controlled through a simple pricing adjustment. Non-separability implies very high costs of finding prices, to cite one of Coase (1937) category of transaction costs. 1.2. Related literature Dual distribution, where the franchisor simultaneously runs company-owned and franchised outlets, should be seen as part of the choice between franchising and running company stores. A number of efficiency explanations of dual distribution have been suggested based on monitoring costs (Rubin, 1978), signaling of reputation (Gallini & Lutz, 1992; Lafontaine, 1993), and search costs for markets (Minkler, 1992). Capital constraints are also cited (Caves & Murphy, 1976).2 Some authors take a strategic approach to dual distribution, focusing on interactions among rivals in the product market (Slade, 1998a and Slade, 1998b). Vertical integration is based on strategic motives. For example, Slade (1998a) shows that when an improvement in the price/cost margin can be achieved with delegation, the price decision is decentralized to local retailers favoring franchising. Conversely, when delegation does not improve the margin, the price decision is taken by the manufacturer favoring company-owned outlets. Bai and Tao (2000) have suggested that a strategic explanation of dual distribution has to do with brand development and sales activity, franchising being chosen when the latter is more important and company-owned outlets being preferred when the first is more relevant. Bradach (1997) finds that a company-owned network or a completely franchised organization carries its own advantages and weaknesses, but dual distribution will be better able to maintain uniformity and the achievement of system-wide adaptation to changing markets. No single explanation of the choice between franchising and integration, including the selection of dual distribution, is likely to work for all observed systems at all times. We offer the externality-based explanation that follows in this spirit. Surely we cannot tell the type of externality story we have in mind at all types of franchises and dual distribution. A further caution: we do not claim that integration must occur in the face of non-separability, e.g. imperfect externality control could be preferable given that much better incentives to cost control are preserved under franchising. Also, if externalities are insignificant, we should expect independent firms to operate the distribution system since there would seem to be no gains from trade from any form of coalition formation in such cases. Non-separability is neither a sufficient nor a necessary condition for merger. Nonetheless, it is an important possible incentive towards integration, likely to be of empirical relevance. Our theory refers to a trade off between managerial motivation (including local demand conditions) and externality control. The parent company's cost of controlling externalities within the entire system varies with the number of outlets of each type. The externality can flow between franchised outlets, between company-owned outlets and between franchised and company-owned outlets in either direction. One possibility is to think of the externality as something like an outlet's cost of free servicing attached to a product sold elsewhere in the system. It may be desirable to combine the initial product sale with widely available free servicing to encourage proper servicing and create a reputation for reliability. The parent company has to find a pricing or administrative mechanism to obtain the system-wide cooperation as an outlet's external cost is internal to the system, and it wishes to offer a service network for which customers have paid. It is more costly to control the externality the greater the number of either type of store. We assume it is harder to gain the cooperation of a franchisee over controlling externalities between outlets than it is to gain the cooperation of a manager of a company store, broadly because the manager is subject to incentives of promotion or job tenure, whereas the franchisee is seeking to avoid costs. Therefore, the addition of a franchisee increases externality control costs by more than is the case from the addition of a company store. The trade off could also emerge because it is more difficult to motivate managers of company-owned outlets than franchisees as far as local business efficiency is concerned. A dominant view in the literature is that local motivation over such things as cost minimization or finding new business is stronger for franchisees than for managers. We suggest that the local superiority of the franchisee arises because business decisions can be made without constant reference to superiors, which is unlikely to be the case for a hired manager. This is why franchising is not quite the same thing as hiring profit-sharing managers.3 Cost minimization may also be better where the local franchisee has a freer hand to try techniques (Minkler, 1992). The same freedom that gives a strong motivation to local profitability causes a problem with externality control if the local incentive is to free ride. In our analysis, when the externality is separable, franchising is preferred to company-owned outlets as long as it is easier to motivate franchisees rather than managers. Dual distribution emerges as optimal when externality control becomes more important.

Our paper shows that the nature of externality has important organizational consequences. Factors raising the costs of internalizing externalities affecting parts of a network business make integration more attractive than franchising. Non-separability of the externality is just such an effect. We have shown that an incentive to integrate can arise from the very nature of the externality itself. Dual distribution arises naturally as a consequence of the comparison between franchising and integration. The empirical application of our model should be fruitful in cases where strong systemic effects exist within franchise networks. Proxy variables for systemic effects comprise such things as service requirements for products, perhaps as indicated by the existence of warranties. We suggest that econometric modeling may be improved by including servicing commitments alongside factors like the geographical density of outlets as explanatory variables for the proportion of franchised outlets. Our paper also shows that franchising is not a universally applicable organizational form. It will not work if the advertising charges or transfer fees cannot be calculated in terms of a simple, low-cost rule for franchisees, which is likely to be difficult in the case of non-separability. Non-separability can make the costs of coordinating the franchise network very high. Merger is more likely when non-separable externalities are significant relative to the costs of their control.