The sample for this study consisted of 157 new franchise systems established in the United States in 1981, 1982 or 1983; identified from the Sourcebook of Franchise Opportunities, 1985 (Bond, 1985). This source has been used in several studies of franchising (Baucus et al, 1993; Sen, 1993) and has been found to provide an unbiased sample of franchisers. The 1985 issue was selected since the data for this issue was gathered in 1984 about franchise systems established through December 31, 1983. New franchisers were those franchise systems which Sourcebook of Franchise Opportunities, 1985 indicated first began to franchise in 1981, 1982 or 1983. The firms were drawn from a number of industries.

Dependent Variable

Shane (forthcoming) found that three quarters of all new franchise systems failed within ten years of their formation. Carney and Gedajlovic (1991) and LaFontaine and Kauffman (1994) have argued that the high rate of failure of new franchise systems suggests that the survival of a new franchise system over time is an important measure of performance. Therefore, survival of the franchise system was used as the dependent variable in this study.

In this study, survival was operationalized as the existence of the franchise system in 1994. This condition was demonstrated through a three-step process. First, franchise systems were categorized as surviving if they were listed in the Sourcebook of Franchise Opportunities, 1994 (Bond, 1995). Second, if firms were not listed in the Sourcebook, they were contacted by telephone at their last known telephone number. If the firms were contactable, they were asked if their franchise system was still in operation. Third, if the firms were not contactable, or were contacted and found to no longer be franchising, the franchise system was categorized as non-surviving.

It is important to note that this technique allows one to capture firms that had been acquired or changed names. These firms indicated such changes when I contacted them by telephone. Moreover, this technique is appropriate for testing the survival of franchise systems. Since franchisers need to attract franchisees to have viable franchise systems, surviving franchise systems have a strong incentive to be contactable. Therefore, the inability to contact a franchiser is a good indication that the franchise system is no longer in existence.

Independent Variables

This study employed a number of independent variables derived from agency theory that should influence the survival of the new franchise systems. These were: passive ownership, monitoring efficiency, geographic dispersion, royalty rate, master franchise agreements, franchise fees, contract length, and complexity of the franchise concept.

Passive ownership is ownership of the retail outlet by someone other than an owner-operator. It inhibits the survival of new franchise systems by removing ownership incentives, which are one of the primary motivations for the use of hybrid organizational forms. Passive ownership is measured by a dummy variable in which one indicates that the franchiser uses passive ownership, following the model of Michael (1992).

Monitoring efficiency is a measure of the cost of the franchiser is adhering to the terms of the franchise agreement. It enhances the survival of new franchise systems by allowing the firm to reduce the cost of deterring agency problems. Monitoring efficiency is measured as the ratio of the number of outlets to franchiser staff. In this ratio, a part-time staff member is counted as one half of a full-time staff member. Outlets per staff member is a good measure of monitoring efficiency because franchisers typically monitor franchisee behavior by having their personnel visit franchisee-owned outlets to verify that the franchisee is adhering to the terms of the franchise agreement.

Geographic dispersion is the degree to which the expansion of the franchise system is concentrated in a certain area. The more geographically dispersed a franchise system, the more costly it is to monitor franchisees. Therefore, geographic dispersion inhibits new franchise system survival. This study measures geographic dispersion as the number of different regions of the United States into which the franchiser was expanding in 1984. Similar measures have been used in studies by Norton (1988a), Brickley and Dark (1987) and LaFontaine (1992).

The royalty rate is the ongoing percentage of sales that franchisees pay to the franchiser. The payment of a percentage of franchisee sales to franchisers on an ongoing basis is the best incentive for the franchiser to monitor franchisees. Therefore, the royalty rate enhances new franchise system survival. The royalty rate was measured as the percentage royalty rate stated in each franchiserís Uniform Franchise Offering Circular, following LaFontaine (1992) and Sen (1993).

Master franchise agreements are agreements to allow an organization to sell individual franchises in a particular location. They inhibit the survival of new franchise systems by requiring contractual codification of monitoring behavior, which makes it more difficult to monitor franchisees. The use of master franchise agreements is measured by a dummy variable in which one indicates that the franchiser uses master franchise agreements.

Franchise fees are the up-front fees that franchisees pay to the franchiser upon the signing of the franchise contract. High franchise fees indicate the potential for quasi-rent appropriation and inhibit the survival of new franchise systems by increasing contracting difficulties. The variable is measured as the dollar amount of the franchise fee indicated in each franchiserís Uniform Franchise Offering Circular. This variable has been used in studies by Martin and Justis (1993), Carney and Gedajlovic (1991), Caves and Murphy (1976) and Combs and Castrogiovanni (1994).

Contract length is the number of years for which the franchise agreement is operative. Long-term contracts enhance the survival of new franchise systems by reducing the amount that agents can expect to gain from shirking relative to the amount they can expect to gain from not shirking. This variable is measured in years, following Michael (1992).

Complexity is a measure of how complicated it is to execute a franchise concept. Low complexity enhances the survival of new franchise systems because it makes hybrid arrangements easier to develop and manage. Complexity was measured as a count of the number of different support services that the franchiser contracts to provide to the franchisee as part of the franchising package. Franchise systems with more support services are more complex than are systems with few support services. A similar variable was used by Sen (1993).

Control Variables

The first control variable was the number of years that the firm was in existence when it first offered franchises. This variable was controlled because the vast literature in population ecology has convincingly demonstrated the liabilities of newness on firm survival (Stinchcombe, 1965; Aldrich and Auster, 1986). This variable has been used in franchising studies by Combs and Castrogiovanni (1994), LaFontaine (1992), Martin (1988), and Martin and Justis (1993).

The second set of control variables was a pair of dummy variables for the year that the franchise system was established. One dummy variable was used for franchise systems that began in 1981. The second dummy variable was used for franchise systems that began in 1982. These variables were controlled because the percentage of surviving franchise systems decreases over time. Therefore, franchise systems started in 1981 and 1982 should be less likely to be alive in 1994 than franchise systems started in 1983.

The fourth set of control variables was for industry. This variable was operationalized as a dummy variable for each of the industries for which there were new franchise systems established during the 1981-1983 period, except the business services industry. Industry was defined according to the U.S. Commerce Department's classification of franchisers as shown in the Sourcebook of Franchise Opportunities, 1985 (Bond, 1985), rather than by SIC code. This classification was used to be consistent with previous research on franchising (e.g., LaFontaine, 1992; Combs and Castrogiovanni, 1994; Brickley and Dark, 1987; Anderson, 1984; Michael, 1993; LaFontaine and Kauffman, 1994; Martin, 1988).

Industry was controlled for several reasons. First, industries vary on the complexity of franchise concepts (Michael, forthcoming). Second, the value of the brand name and the incentive to free ride vary across industries (Fladmoe-Lindquist and Jacques, forthcoming). Third, the labor intensity of an industry affects the likelihood of agency problems (Norton, 1988a). Fourth, the use of monitoring and royalty payments in franchise arrangements varies across industries (Lal, 1990). Fifth, the ability of the franchiser to enforce contractual provisions on the behavior of franchisees varies by industry (Caves and Murphy, 1976).


A logistic regression model that predicted survival of the franchise system was used. The logistic regression measures the effect of the independent variables, estimated in 1984, on the likelihood that the franchise system would be operating in 1994. A positive coefficient demonstrates that the independent variable has a positive effect on survival.

Top of Page
Previous Page Main Page Next Page
Return to 1996 Topical Index

© 1997 Babson College All Rights Reserved
Last Updated 4/24/97 by Jo Ann Mathieu

To sign-up for the Center for Entrepreneurial Studies' publication lists,
please register with the
Entrepreneurship WebTeam.