منطق قراردادهای فرانشیز (فرانچایز) : نتایج تجربی از ژاپن

This paper examines the factors determining the choice between company-owned outlets and franchising. A number of researchers have based their studies on the data of retail contracting in the US. However, there has been little empirical analysis in the case of Japan. In this paper, we provide a comprehensive assessment of various alternative hypotheses about franchising by using Japanese data. The principal-agent model of franchising that assumes risk-sharing and moral-hazard is supported by our data.

In recent years, many firms have adopted franchising as a method of doing business in a variety of industries. We can easily find franchising in the restaurant (fast-food) sector which includes such well known chains as McDonald's, KFC, and Pizza Hut among others, but it has also steadily grown in importance in the non-food retailing sector. This is especially true in Japan as evidenced by the fact that the number of convenience stores operating under franchise arrangements has rapidly grown since the 1970s. In 2007, Seven-Eleven Japan had franchised 93.3% of its 12,034 stores, Lawson had franchised 94.9% of its 8587 stores, and FamilyMart had franchised 93.1% of its 6691 stores. Seven-Eleven Japan has had the largest sales of all retail firms in Japan since 2001. On an all-store basis, in 2008, the sales of convenience stores in Japan totalled 7.86 trillion yen, which is more than the department store sales of 7.38 trillion yen. Hence, it is timely to examine the determining factors behind franchising. Why have so many firms chosen to franchise rather than expand through company-owned outlets? A variety of factors have been shown to have significant impacts on the franchising decision.1Rubin (1978) first explained the choice of franchising in terms of monitoring and control within the system, drawing on the agency analysis. There have been theoretical developments in this area.2 It has been explored that the problem of moral-hazard on the part of both the franchisor and the franchisee is essential to the specification of franchise contracts. In the principal-agent model, franchising (vertical separation) provides higher performance incentives than company-owned outlets (vertical integration) in which employees are paid a set wage not closely tied to observed performance, but also subjects agents to greater risk, which is costly. Where heightened incentives are more valuable, or their associated risk-bearing costs are lower, franchising is more profitable and more likely. On the empirical side, a number of authors have made efforts to test theoretical hypotheses about the franchisor's monitoring costs of franchisee efforts,3 the shedding or sharing of risk involved in the operation of outlets,4 franchisee moral-hazard,5 and franchisor moral-hazard.6 The literature gives a number of other explanations for franchising such as, raising capital for expansion (Caves and Murphy, 1976), signaling of business quality (Gallini and Lutz, 1992 and Lafontaine, 1993), and free-rider issues for franchisees (Brickley and Dark, 1987 and Norton, 1988). Lafontaine (1992) uses a large sample of franchisors to provide an assessment of various hypotheses on franchising, including risk-sharing, and both franchisee and franchisor moral-hazard. Much of the existing empirical analysis has been done using North American data; however, franchising has also developed in other countries, including Japan. Hence, it is necessary to provide a comparative work that assesses the robustness of the empirical results in other geographic areas. The purpose of this paper is thus to extend empirical work to Japan. First, we attempt to test a similar empirical model with Lafontaine (1992), using our Japanese data. Next, we add two hypotheses on signaling and free-riding to the original model. The contribution of this paper lies in its attempt for providing a comprehensive empirical assessment of various alternative hypotheses about Japanese franchising, which has not previously been studied. We employ firm-level data for 572 franchisors involved in 33 industry sectors in Japan in 2002, including retailing (convenience stores, supermarkets, medicines/cosmetics, etc.), services (dry cleaning, hair salon/health and beauty, real estate agencies, etc.), and restaurants. This was obtained from the annual publication Nihon no Franchise Chain 2003 (Japanese Franchise Chain, 2003) published by Shogyokai. The data was collected through a questionnaire mailed to franchise chains by Shogyokai. The main results obtained from our empirical analysis are as follows. First, the agency-theoretic model that assumes risk-sharing and double-sided moral-hazard is supported by our data. Second, the traditional capital explanation for franchising is not empirically supported by our data. Those are consistent with previous results based on the US data. Third, factors such as signaling of business quality and franchisee free-riding significantly affect franchisor decisions to use franchising. This is interesting because previous empirical studies do not support these hypotheses. The rest of the paper is organized as follows. Section 2 describes some characteristics of franchising in Japan. Section 3 discusses the literature and hypotheses. Section 4 details the data and variables for the study. In Section 5, we show our methodology and results. We conclude our study in Section 6.

This paper extends empirical work on the determinants of franchising using Japanese data. Since there has been little empirical work on Japanese franchise activity, this would be a welcome addition to the literature. The traditional capital explanation for franchising is not empirically supported by our data. In contrast, we show results supporting the hypotheses of risk and double-sided moral-hazard. We add hypotheses of signaling and free-riding in this paper, which are both statistically supported. Therefore, our results obtained from Japan support a principal-agent model of franchising. This is the most important contribution of this paper vis-à-vis the existing literature. In this paper, we have provided an empirical study on the franchisors’ decisions to adjust the agency problems by varying the extent to which they use franchising and by opting for different contract terms regarding franchise fees and royalty rates. At a theoretical level, it would be interesting for future research to study the relationship between the choices of ownership structure and fee structures in more detail.