New Firm Survival: Institutional Explanations for New Franchisor Mortality *
Scott Shane, Massachusetts Institute of Technology
Maw-Der Foo, Massachusetts Institute of Technology and National University of Singapore
Abstract
Introduction
Business Format Franchising in the
United States
Theory Development and Hypotheses
Methodology
Results
Discussion
Conclusion
Table 1
Notes
References
Why do some new firms succeed and others fail? Economists argue that new firms fail because entrepreneurs inefficiently manage production and organizational design (Williamson, 1985). Sociologists (e.g., Granovetter, 1985) have typically viewed this explanation as undersocialized, and argue that institutional legitimacy must also be considered to explain the survival of new firms. This paper examines the survival of 1292 new franchisors established in the United States from 1979–1996. The results show that institutional legitimacy adds to economic explanations for the survival of new franchisors and suggests the importance of a properly socialized explanation.
Economists typically argue that the survival of new firms depends on the efficiency of their production processes and their organizational designs (Williamson, 1985). When examining franchisors, economists have argued that survival depends on their ability to achieve efficiencies in three areas: production, resource acquisition, and contracting. First, franchise systems exploit scale economies in marketing, production, and administration (Caves and Murphy, 1976). New franchisors which are established below minimum efficient scale must grow rapidly in the early years for the franchisor to survive. Second, franchising is an efficient way to obtain labor and capital (Lafontaine and Kaufmann, 1994). When new franchisors are growing rapidly or need to obtain large amounts of capital, the relatively greater use of franchising will be survival enhancing (Shane, 1996). Third, franchisors which economize on the agency costs of adverse selection, moral hazard, and hold-up, and engage in efficient risk bearing between the principal and the agent should be more likely to survive over time (Fama and Jensen, 1983).
Sociologists (e.g., Granovetter, 1985) have criticized this explanation as under-socialized. They argue that the success of new firms depends not only on economic efficiency, but also on institutional approval (Hannan and Freeman, 1984). In particular, they argue that firm survival depends on the ability to establish cognitive and socio-political legitimacy (Aldrich and Fiol, 1994). Moreover, they hold that policies adopted at the time of founding imprint the firm with the characteristics of that time and place, which continue to influence the survival of the firm over time (Stinchcombe, 1965).
In this study, we explore whether institutional theory adds to economic explanations for the survival of new franchise systems. In specific, we examine the effect of cognitive legitimation, sociopolitical legitimation, and imprinting on the survival of 1292 new business format franchisors established in the United States between 1979–1996, controlling for factors found to be important in previous studies.
This study makes three important contributions. First, we show that institutional theory adds to economic explanations for the survival of new franchisors, suggesting that economic explanations alone are undersocialized (Granovetter, 1985). Second, the study contributes to institutional theory by demonstrating the survival value of certification in a multi-industry context. Third, the study contributes to our understanding of franchising by correcting erroneous conclusions of prior cross-sectional research, through a research design that avoids the problems of path dependence, censoring, and selection bias.
Business Format Franchising in the United States
Taking place in over fifty industries, ranging from banking to the internet, business format franchising is a mode of distribution of goods and services by which an individual (the franchisee) obtains from another individual (the franchisor) the rights to use a trade name and operating system in return for oversight by the franchisor and the payment of royalties. While the origins of franchising in the United States can be traced to the Singer Sewing Machine Company in the mid 1850’s, the current, regulated era of franchising can be traced to October 21, 1979. On that date, the Federal Trade Commission issued a trade regulation rule, which required franchisors to disclose to prospective franchisees specific information about the franchisor organization, its principals and the investment that the franchisee must make to enter the franchise system. In this study, we examine the survival of new franchisors established during the current regulated period.
Theory Development and Hypotheses
Institutional theory argues that firm survival depends, in part, on the acquisition of cognitive and socio-political legitimacy (Aldrich and Fiol, 1994). Cognitive legitimacy is the degree to which an organization’s activities are taken for granted. Socio-political legitimacy is the “extent to which a new form conforms to recognized principles or accepted rules and standards (Aldrich and Fiol, 1994: 646).” Legitimacy enhances survival by making it easier for new firms to obtain access to resources, attract customers (Wiewel and Hunter, 1985), answer challenges about competence, combat competitive, and achieve perceived reliability (Hannan and Freeman, 1984). Institutional theory also argues that policies established at founding imprint a firm with the characteristics of that place, and that these policies influence later firm survival (Stinchcombe, 1965). In the section below, we develop specific hypotheses about the effect of cognitive legitimacy, socio-political legitimacy, and imprinting on the survival of new franchisors over time.
Cognitive Legitimacy
When a firm engages in a new activity for the first time, it needs to establish internal and external norms, new roles for organization members, standard operating procedures, and new patterns for interacting. The fact that these activities are not yet taken for granted creates a liability of newness (Stinchcombe, 1965). The organization must expend time and effort for members to learn new roles, socialize strangers, and establish routines and procedures (Hannan and Freeman, 1984). However, over time, people learn routines, procedures, roles, and patterns for interacting in ways that become reliable, reproducible and taken for granted (Stinchcombe, 1965). The taken for grantedness of activities allows the organization to “conserve time and other organizing resources (Aldrich and Fiol, 1994: 648),” which enhance survival prospects. Previous research has shown that cognitive legitimacy is enhanced as firms become older (Carroll and Delacroix, 1982). This argument leads to the first hypothesis:
H1: Age will increase the probability of new franchisor survival.
Firms also have greater cognitive legitimacy when they are. Larger organizations are more visible, more powerful and more prestigious, which enhance their taken for granted nature. Larger firms are perceived as having greater ease in raising capital, having greater long term stability, and as having better internal labor markets (Singh and Lumsden, 1990). Previous research has show that the cognitive legitimacy of firms is enhanced if they are larger (Carroll and Delacroix, 1982). This argument leads to the second hypothesis:
H2: Size will increase the probability of new franchisor survival.
Socio-Political Legitimacy
When a firm engages in an activity for the first time, its management can rarely convince others that it knows the “right” way to do things. Uncertainty about the value of the new organization’s way of doing things makes it difficult to gain the support of stakeholders. The firm can reduce the uncertainty to external constituents by doing things in a way that is already accepted by them as appropriate and valid (Aldrich and Fiol, 1994). One of the most important mechanisms for obtaining socio-political legitimacy is certification by powerful institutional actors. Podolny (1994) explains that when the attributes of an organization cannot be directly observed, external constituents infer them from attributions made by respected institutions. Therefore, certification by institutions that possess social acceptance also can enhance the prospects of a new organization.
Although socio-political approval is granted by many institutions—governments, financial institutions, and other societal actors—certification by the media, which are looked to provide information about new organizations, play a particularly important legitimizing role (Rao, 1994). He explains that publications like A.M Best, Moody’s, Consumer Reports, J.D. Powers, and business magazines like Business Week and Entrepreneur Magazine, legitimate new firms through rankings. Media certification reduces uncertainty, the cost of consumer search, and the difficulty of measuring intangible capabilities, thereby allowing the firms to enhance their perceived status. Moreover, media rankings differentiate the amount of approval allotted to individual firms, allowing different firms to be allocated different levels of approval (Hybels, 1995). Rao (1994:32) has shown that certification is survival enhancing for new firms as it legitimates organizations “because of the taken for granted axiom that winners are ‘better’ than losers and the belief that contests embody the idea of rational and impartial testing . . . [This generates] status orderings of organizations that determine their access to resources [and] enable higher status firms to extract greater rewards for producing even the same good as lower status firms” This argument leads hypothesis 3:
H3: External certification will increase the probability of new franchisor survival.
Imprinting
Firms are imprinted by their place of founding with a particular way of operating that influences the firm even when the initial environmental effect is gone (Stinchcombe, 1965; Carroll and Delacroix, 1982). Dimaggio and Powell (1983) argue that this imprinting occurs because new organizations must adhere to local norms, legal and regulatory policies to obtain resources. Once a policy is established, however, it becomes difficult to change. People prefer predictability and reliability, leading organizations to favor stability (Singh et al, 1986). Moreover, policies become embedded in ties to the institutional environment (Granovetter, 1985); and routines and procedures become perceived as the only acceptable way of doing things. Thus, firms are severely constrained in their ability to change policies previously established (Hannan and Freeman, 1984). In franchising, Lafontaine and Shaw (1996) have observed that once franchisors set contract terms, they change them very little over time. Shane (forthcoming) showed that these policies persist even when they lead to the failure of the franchise system.
The legal environments of the states in which franchisors are founded contribute to this imprinting effect. New franchisors must attract franchisees, but the threat of franchisor hold-up makes this difficult. Franchisees are concerned that franchisors will appropriate quasi-rents by opportunistically threatening to terminate them after they have made specific investments. Therefore, franchisors develop policies to overcome this problem; and the different legal environments in which systems are founded influence these policies. Fourteen states (Arkansas, California, Connecticut, Delaware, Hawaii, Illinois, Indiana, Michigan, Minnesota, Nebraska, New Jersey, Virginia, Washington, and Wisconsin), have termination laws in which franchisors are required to show “good cause” to terminate franchisees before the expiration dates of their contracts (Brickley et al, 1991). “Good cause” is defined as non-compliance with the terms of the franchise contract and does not include general poor performance (Williams, forthcoming). If franchisors terminate franchisees without “good cause” in “termination” states, they are required to compensate franchisees for their losses. Therefore, termination laws reduce the threat of opportunistic hold-up by franchisors by making termination more costly.
By providing franchisees with protection against opportunistic franchisors, termination laws lower the cost of attracting qualified franchisees. Given the greater barriers to termination in termination states, franchisors founded in these states more easily attract applications from qualified franchisees. Consequently, franchises founded in these states are structured to use franchising more heavily relative to company-ownership of outlets. Franchisors established in termination states become imprinted with routines for managing a higher proportion of franchised outlets than firms which are founded in non-termination states. Even when these franchisors expand into non-termination states, and the laws applicable to franchisee termination are different, the initial imprinting of the system’s founding location influences its approach to franchising. Therefore, franchisors founded in termination states are more likely to survive if they maintain a higher proportion of franchised outlets. This argument leads to hypothesis 4:
H4: Franchisors founded in termination states will be more likely to survive if they use franchising more intensely.
Imprinting also influences franchisor policies toward outlet capitalization. The ability of new franchisors to attract franchisees is also constrained by a lack of information about the quality of the new franchise system. This situation creates the potential for low quality franchisors to opportunistically misrepresent their quality to franchisees, and makes franchisees reluctant to contract with franchisors. To combat this adverse selection problem, qualified franchisors adopt policies to disclose information to franchisees. The policies that they adopt are influenced by the legal environments of the states in which the systems are founded. Sixteen states (California, Florida, Hawaii, Illinois, Indiana, Maryland, Michigan, Minnesota, New York, North Dakota, Oregon, Rhode Island, South Dakota, Virginia, Washington, and Wisconsin) have laws which require franchise system registration with state authorities. Registration provides information about the franchisor’s system which reduces the adverse selection problem and makes it easier for franchisors to attract franchisees.
However, registration also influences franchisors to adopt a policy of lower outlet capitalization. Regulators in registration states typically argue that an undercapitalized franchisor is an unreasonable risk to franchisees and deny registration or impose escrow or bonding requirements on poorly capitalized franchisors (Kaufmann, 1992). Moreover, regulators in these states usually allow registration of only those franchise systems for which the franchisor’s net worth exceeds the franchisee’s required capitalization. This preference for lower outlet capitalization among regulators in registration states means that franchisors founded in registration states are designed with lower outlet capitalization. Franchisors founded in registration states become imprinted with low capitalization policies; and their survival depends on the maintenance of these policies. Even as these franchise systems expand to states which do not have registration requirements, these firms are more likely to survive if they have lower outlet capitalization. This argument leads to hypothesis 5:
H5: Franchisors founded in registration states will be less likely to survive if they have high levels of outlet capitalization.
Data
We examined the survival of 1292 new U.S. headquartered business format franchisors established in the United States between 1979 and 1996. The analysis used data collected from Entrepreneur Magazine’s annual listing of franchise systems, which assembles information on the existence or non-existence of the franchisors, the location of the firm at founding, and annual data on each system’s franchise fee, royalty rate, outlet capitalization, franchisee financing, number of company-owned and franchised outlets, and a numerical ranking of the system based on a proprietary formula.
Entrepreneur Magazine obtains its data by surveying annually all franchisors that it knows to exist. Similar to surveys conducted by academics, Entrepreneur Magazine’s list only includes information from firms which choose to respond to the survey. Although no source identifies the entire population of business format franchisors operating in the United States, previous researchers have estimated Entrepreneur Magazine captures half of all franchisors and most new ones (Lafontaine, 1995). Entrepreneur Magazine verifies the information contained in the magazine; and franchisors have a strong incentive to provide accurate information since prospective franchisees are likely to verify it (Scott, 1995).
The dependent variable was the probability of franchisor failure. We define entry as occurring in the year that Entrepreneur Magazine indicated that the franchise system was established. We define failure as exit from Entrepreneur Magazine’s listing, and non-reemergence in the listing at a later date. Previous research has found that exit from Entrepreneur Magazine accurately represents exit from franchising. In a study of the entry and exit of new franchisors over the 1980–1992 period, Lafontaine and Shaw (1998) compared the exit from Entrepreneur Magazine with exit from another trade publication, Franchise Annual. Using the same data as is used in this study, but for the 1980–1992 period, these authors found that exit from Entrepreneur Magazine accurately measured exit from franchising. Similarly, Shane (1996:224) examined 138 firms that first began to franchise in the United States in 1983 and found that none of the firms in his sample that were delisted from Entrepreneur Magazine, “show up in later years of Entrepreneur Magazine or Franchise Annual.” Lafontaine and Shaw (1998) conclude that the Entrepreneur surveys represent quite well the amount of entry and exit into franchising.
During the 1979–1996 period, 1097 (85 percent) of the franchisors failed. These failure rates are consistent with Shane (1996) who found a 75 percent failure rate of new franchisors across a ten year period, and Lafontaine and Shaw (1998) who found a 70 percent failure rate over twelve years. Analysis was conducted on 1292 firms, called cases by researchers of firm mortality. Cox regression analysis was used to test the effects of the covariates on franchisor survival. Cox regression analysis controls for the effects of right censoring, and thus avoids the problem of biased estimators that exists when a case is misclassified as surviving when, in fact, it has only not been observed to fail. Previous organizational ecology research (Hannan and Freeman, 1984) shows how right censoring leads to biased results with alternative statistical techniques, such as logistic regression.
The failure rate model was estimated as an instantaneous hazard rate. Each franchise system’s information was broken into annual spells, with the new covariate information included each year. There are 3608 firm-year observations 1 in the sample. A firm year observation is the set of covariate values for a particular firm for a particular year. The effects of the covariates were tested against survival during the following calendar year. As is the norm in research on firm mortality, firms that failed during the observation period were labeled events. If the organization did not fail during the period of the study, it was treated as right censored.
Operationalization of the Institutional Theory Covariates
Following institutional theory arguments about cognitive legitimacy (Aldrich and Fiol, 1994), we test the effects of firm age and firm size on firm survival. Age is operationalized as the log of the number of years since firm incorporation. Size is operationalized as the log of the number of units in the chain (Ingram and Baum, 1997).
Following Rao (1994), we test the effects of media certification by operationalizing Entrepreneur Magazine’s ranking of franchise systems. The ranking is based on a proprietary formula that incorporates the quantitative data that is reported in the magazine with the subjective assessment of the magazine’s experts. By incorporating the objective measures into the regression models independently, our measure of Rank represents the behavioral response of people to the system and measures certification as defined by institutional theory.
We recognize that the Entrepreneur Magazine ranking is an imperfect measure. Some researchers question the use of rankings as a measure of certification because rankings envision a behavioral response of people to certification itself and incorporate many different reputations including those of financial stability and strength, litigation history, and termination policy. Moreover, to be listed in Entrepreneur Magazine, a franchise system must respond to the magazine’s survey; and to be ranked the franchisor must provide information that the magazine can verify to be accurate by comparison to a uniform franchise offering circular. Despite these weaknesses, institutional theory argues that rankings, such as those of Entrepreneur Magazine, are valid measures of certification because media rankings represent the overall reputation of a firm, and because people act in response to this combined reputation when making decisions (Rao, 1994). Since previous research on certification (e.g., Rao, 1994) showed that the best functional form for the rank variable is its logarithm, we operationalize Rank as the log of the annual ranking of the franchise system by franchising experts in Entrepreneur Magazine. We measure the effects of imprinting by examining the effect on system survival of the interaction of franchisor policies with founding in different state legal environments. To capture the effects of state regulation, two dummy variables were created. Termination was coded one if the franchise system was founded in a state in which franchisee termination laws exist and zero otherwise. Similarly, Registration was coded one for franchise system was founded in a state in which registration requirements exist and zero otherwise.2
Operationalization of the Economic Theory Controls
We control for several variables taken from economic theory. First, we operationalize System Growth as the percentage rate of growth in the number of outlets over the previous two years.3 Firm growth enhances the survival of new franchise systems by enabling them to reach minimum efficient scale (MES) more quickly. Since scale economies in marketing, purchasing and administration reduce per unit costs as size increases in many industries in which franchising takes place, these economies provide an efficiency advantage to size by lowering costs relative to those of competitors. Therefore, new firms need to grow quickly or they will be driven out of business by more efficient competitors who operate with lower per unit costs.
Second, we control for the system royalty rate because it provides the primary incentive to both franchisors and franchisees (Lafontaine, 1992). Franchisors with too high royalties have franchisee incentive problems; while franchisors with too low royalties have franchisor incentive problems. We do not predict the direction of this effect since different researchers have argued that royalty rates should have a positive, a negative or a curvilinear effect (Lafontaine and Shaw, 1996). Royalty Rate is operationalized as the ongoing percentage of sales that franchisees pay to the franchisor for the use of the trademark and operating support. Where a range of royalties was reported, the average was used. Where a flat royalty was reported, it was divided by the industry average level of sales to create a percentage.
Third, we control for the franchise fee, which is an investment in system specific assets which generates quasi-rents, motivates hold-up and makes the franchisee reluctant to invest (Lafontaine, 1992). Franchise Fee is operationalized as the dollar value of the up-front fee that the franchisee pays to the franchisor to purchase an outlet. Where a range was provided, we used the mean.
Fourth, we include a dummy variable if the franchisor provides financing to the franchisee to control for franchisor resource constraints (Lafontaine, 1992). Firms which have sufficient capital to finance franchisees are more likely to have the necessary resources to survive over time (Caves and Murphy, 1976).
Fifth, we examine the impact of efficient contracting for resources. We operationalize the relative emphasis of the franchise system on company-ownership of outlets as Company, the percentage of total outlets that are company-owned. We examine the interaction of Company X Capitalization (the amount of money needed to open an outlet). Agency theory arguments about efficient risk bearing suggest that firms which have a high level of capitalization and a large percentage of company-owned outlets should be less likely to fail. When the agent is more risk averse than the principal, it is more efficient for the principal to insure the agent. Franchisees are typically more risk averse than franchisors because franchisees’ risk involves employment risk, which cannot be diversified and because franchisors typically have more wealth than franchisees (Brickley and Dark, 1987). When risk averse agents are required to bear risk, they make inferior investment decisions than less risk averse actors. Since agent risk is a function of the size of the investment, inefficient risk bearing increases with the size of the investment that the franchisee is required to make (Brickley et al, 1991). Therefore, as the required investment by franchisees increases, franchisors that design agency contracts to allocate risk to themselves by owning more outlets should be less likely to fail.
Sixth, we examine the interaction between Company X System Growth. Agency theory suggests that firms which are growing rapidly should be more likely to fail if they rely more heavily on company-owned outlets. Franchising reduces the cost of agent selection since investment in a franchise system provides an incentive for agents to self-select. As explained above, firms are more likely to experience adverse selection problems in identifying employees when they grow rapidly. Thus, rapid growth raises the value of mechanisms like franchising that minimize the adverse selection problem by providing an incentive for outlet operators to self-select (Shane, 1996).
Seventh, we examine the interaction of Company X Time. Previous research (e.g., Lafontaine and Kaufmann, 1994) has argued that an emphasis on franchising is advantageous in a system’s early years because the franchisor needs to obtain resources and establish a brand name. These activities are enhanced by franchising, which allows more outlets to be created per unit of resources than does company ownership of outlets.
Other Controls
We also include controls suggested by ecological theory. Ecological theory argues that firms are relatively non-adaptive and suffer from a liability of change which reduces the reliability stakeholders favor (Hannan and Freeman, 1984). We measure Fee Growth, the two year percentage change in the franchise fee; Royalty Growth, the two year percentage change in the royalty rate; Capitalization Growth the two year change in the amount of capitalization necessary to open an outlet.4 We also measure Finance Change as a dummy variable of negative one if the franchisor eliminated financing, zero if the franchisor did not change financing policies and one if the franchisor added financing.
Because the survival prospects of firms vary by industry (Fladmoe-Lindquist and Jacques, 1996), we also used a set of dummy variables for different industry sectors, except for computer-related industries, to control for industry. We control for the rate of growth of gross state product to capture the portion of firm survival that results from between state variation in economic performance, rather than from between state differences in franchise regulation.
Table 1, placed at the end of the paper, shows the Cox regressions predicting franchisor failure. The regressions measure the effect of the independent variables on the likelihood that the franchisor will fail. A positive coefficient demonstrates that the independent variable has a positive effect on franchisor failure.
Model 1 shows the effect of the control variables. Model 2 adds the effect of the institutional theory variables. Model 3 adds the interactions of System Growth X Time and Company X Time. Model 4 tests the robustness of the stock variables, by dropping the change variables and increasing the sample size. Model 5 eliminates the outliers for skewed variables that could not be transformed to test the robustness of the results for these variables.
Model 1 (chi-square = 180.06, p < .05) shows that the economic, ecological and industry control variables explain some of the variance in the survival of new franchisors over time. Among the industry differences, new franchisors in the Amusement (Exp B = 2.86, p < .10), Educational Services (Exp B= 3.71, p < 0.05), Home Furnishings (Exp B = 2.41, p < 0.10), Lodging, (Exp B = 3.44, p < .05), Photographic and Video Services (Exp B = 4.68, p < 0.01), Retail Food (Exp B = 3.41, p < 0.05), Quick Service (Exp B = 2.33, p < 0.05), and Miscellaneous Retail (Exp B = 2.58, p < 0.05) industries, are more likely to fail than new franchisors in other industries. However only the effect for Photographic and Video Services is robust across all models. Among the ecological controls, Financing Change (Exp B = 1.69, p < 0.01) and Fee Growth (Exp B = 1.0a, p < 0.01) increase the probability of system failure; however, the effects of these variables are not robust across all models.
Among the economic theory variables, Company X System Growth (Exp B = 1.01, p < 0.01), makes the new franchise system more likely to fail. Holding Company at the mean value, we find that the proportional hazard for failure is 34.31 percent when System Growth is in the lower quartile and 97.49 percent when System Growth is in the upper quartile. In Model 1, Capitalization X Company (Exp B = 0.99, p < 0.10) makes the new franchisor less likely to fail.
However, when we add the institutional theory variables in models 2,3 and 5, we find that the effect of effect of Capitalization X Company on firm failure turns positive. For example, in Model 2, when we hold Company at the mean value, we find that the proportional hazard for failure is –27.07 percent when Capitalization is in the lower quartile and 5.16 percent when Capitalization is in the upper quartile. Since much of the prior agency theory work on franchising fails to control for firm age and size (e.g., Brickley and Dark, 1987; Brickley et al, 1991), our results suggest this prior work is underspecified. While contradicting the efficient risk bearing hypothesis, the results for this variable support Shane (1996), who argued that efficient contracting for resource acquisition purposes enhances new franchisor survival.
Model 2 adds the institutional variables (chi-square = 415.51, p < .01). Age (Exp B = 0.39, p < 0.01), Size (Exp B = 0.62, p < 0.01), and Rank (Exp B = 0.99, p < 0.01), significantly reduce the probability of new franchisor failure. As was expected, Company X Termination (Exp B = 1.02, p < .01), increases the probability of new franchisor failure. For a termination state, we find that the proportional hazard for failure is 2.25 percent when Company is in the lower quartile and 23.17 percent when Company is in the upper quartile. Also as expected, high capitalization franchise systems had higher failure rates if headquartered in registration states, Capitalization X Registration (Exp B = 1.16, p < 0.10). For a registration state, we find that the proportional hazard for failure is 53.57 percent when Capital is in the lower quartile and 58.04 percent when Capital is in the upper quartile.
Model 3 (chi-square of change = 441.21, p < 0.01) examines the differential effect of System Growth and Company on the survival of new franchisors over time. The main effect of System Growth reduces system failure (Exp. B = 0.99, p < .01), but this effect dissipates over time, as shown by the interaction of System Growth X Time (Exp B = 1.01, p < .01). Holding System Growth at the mean value, we find that the proportional hazard for failure is –48.63 percent when Time is in the lower quartile and –40.10 percent when Time is in the upper quartile. Company X Time has no significant effect on system failure.
Model 4 5 (chi-square = 850.68, p < 0.01) examines the robustness of the model, by eliminating the change variables. This allows us to increase the sample size to include those franchisors which failed in their first two years and tests the consistency of the effects of the remaining variables to the group. All of the hypothesized variables remain significant and signed in the same direction, demonstrating the robustness of the results for these variables. Model 5 (Chi-square = 427.4, p < 0.01) examines the robustness of the model by eliminating outliers for skewed variables that could not be transformed. 6 The results suggest that the results are not driven by outliers and are robust.
This study was motivated by the need to link economic and sociological explanations for the survival of new franchisors. While economists argue that efficiency is an important determinant of the survival of new organizations (Williamson, 1985), sociologists argue that obtaining legitimacy is central to this process (Hannan and Freeman, 1984).
This study demonstrates the survival of new franchise systems is better explained by adding institutional explanations to economic ones. After controlling for economic and ecological factors influencing the survival of new firms, we found support for our five hypotheses. Firm age (hypothesis 1), firm size (hypothesis 2), and media certification (hypothesis 3) reduced the hazard of failure. New franchisors were also imprinted by the location of their founding. Franchisors founded in termination states were less likely to fail if they made more intensive use of franchising (hypothesis 4). Franchisors founded in registration states were less likely to fail if they had low levels of outlet capitalization (hypothesis 5). These results suggest that by modeling the survival of new firms as a function of both economic efficiency and institutional factors, researchers can develop explanations for the survival of these firms that are neither over- nor undersocialized (Granovetter, 1985).
Institutional Theory
This study also contributes directly to the development of institutional theory by providing evidence for the effects of certification on modern firms. While Rao (1994) provides important empirical evidence for the survival value of certification, fundamental economic and social changes since the turn of the century raise questions about the validity of his findings to the current era. Moreover, this study improves upon Rao (1994) by providing evidence of the survival value of certification in a cross-industry study. Previous studies of certification examine its effects only in one industry even though many forms of certification span industries.
Second, although previous research (e.g., Brickley et al., 1991) has shown that the legal environment influences franchisor behavior, the results provide the first evidence that the legal environment in which a franchise system is founded imprints a firm and influences the survival of the franchise system over time. This result is important to policy makers concerned with the impact of legislation on entrepreneurial activity. Moreover, since entrepreneurs may choose the location where they establish a franchise chain, this imprinting effect has implications for both institutional theory and strategic choice theory that future research should explore.
However, this study also indicates that institutional theory’s opposition to the importance of market forces (e.g., Oliver, 1991) is misplaced, at least in the context of a competitive, for profit setting, like franchising. The results here show that economic efficiency does matter in explaining the survival of contractual organizational arrangements.
Economic Theory
The results support some dimensions of economic theory better than others. We find support for Shane’s (1996) argument that the emphasis on contracting is survival enhancing for new franchisors that are growing rapidly or have high capitalization outlets. We also find that growth is survival enhancing in the earliest years of the franchise system, but is survival inhibiting as the franchise system ages. However, royalty rates and franchise fees 7 had little effect on the survival of new franchisors, despite arguments that the royalty rate is the key incentive to both franchisors and franchisees, and arguments that franchise fees represent the net present value of future returns to the franchisee from investment in the franchise system. This finding is important because many economic models of franchising suggest that the formal terms of the franchise contract are central to franchisor and franchisee performance (e.g., Lafontaine, 1992).
Moreover, we find no evidence that increased royalty rates or a shift to company-owned outlets enhance franchise system survival as the system brand name develops as agency theory argues. We also find no evidence that franchisor survival is enhanced by increasing company-owned outlets at the margin as capitalization increases to the potential for franchisee inefficient risk bearing (Brickley and Dark, 1987). In fact, once we control for omitted variable bias, we find that companies that increase franchising at the margin as capitalization increases are more likely to survive.
Finally, we found little support for systematic industry differences in new franchisor survival. Previous research has argued that the survival of new franchise systems should vary by industry because industries vary on the complexity of franchise concepts, the incentive and ability to free ride, labor and capital intensity, the use of monitoring mechanisms, the appropriateness of franchising as a mechanism to pursue business opportunities, and economic attractiveness (Shane, forthcoming. The absence of industry effects is consistent with Lafontaine and Shaw (1998), Shane (1996) and Shane (forthcoming) who found no significant industry effects on franchise system survival rates.
Franchising Research
The results also have implications for franchising research. By examining franchising dynamically, this study avoids the problems of path dependence (Carney and Gedajlovic, 1991) and selection bias (Shane, 1996) that limit the reliability of findings from previous cross-sectional franchising research. For example, the results suggest that dual distribution in franchising (e.g. Bradach and Eccles, 1989) may be an artifact of selection bias. Franchisors typically establish company-owned outlets before they start to franchise and then expand almost exclusively through franchised outlets when they start to franchise. Since franchisors are more likely to survive if they franchise more intensely (Shane, 1996), a cross-section of surviving franchisors will display a mix of company-owned and franchised outlets simply because of selection effects, without adopting a policy of dual distribution.
This study also suggests a methodological direction for future research on franchising. Longitudinal methods are clearly important to franchising research. While publicly available databases on franchising are rather limited, and do not allow for operationalization of many important theoretical constructs, future researchers could develop longitudinal databases by examining government records. For example, franchising researchers could code information contained in Uniform Franchise Offering Circulars archived at state agencies for important system characteristics.
This study is not without its limitations. Since the study uses archival data, several of the constructs were measured by proxy variables. For example, some scholars question the validity of the franchise fee as a measure of hold-up, arguing that these fees represent only a small portion of the franchisee’s specific investment and do not systematically reflect total levels of specific investment (e.g., Bercovitz, 1997). Therefore, the lack of a significant effect for this variable may be a result of a poor proxy rather than a rejection of the underlying argument for hold-up. Nevertheless, the use of proxy variables allowed the examination of the survival of new franchise systems over time and allowed for the examination of the differential effects of franchisor characteristics (such as system growth) over time and provided insights into the dynamics of franchising that would not have been possible with more fine grained, non-archival data. Therefore, readers interested in understanding franchising should examine these results in conjunction with other franchising research that avoid the construct validity questions of proxy variables but lack the ability to examine dynamic research questions. Together these studies will provide convergent validity about franchisor survival.
In conclusion, this study has shown that new franchise systems are more likely to survive if they both gain legitimacy as well as efficiency. This finding expands the franchising literature by demonstrating that sociological explanations enhance economic explanations for the survival of new franchise systems.
1. We use the term “firm-year” to be consistent with previous research on organizational mortality.
2. The current location of the headquarters and the location of headquarters at franchise system founding are statistically indistinguishable in this sample since their correlation exceeds 0.99.
3. The results do not change if one-year system growth is used.
4. Operationalization of these variables as one year growth does not change the results.
5. To conserve space, Models 4 and 5 are not presented, but are available directly from the authors.
6. We could not transform four skewed variables: system growth, fee growth, royalty growth and capitalization growth. Log(10), natural log, and square root transformations for these variables were collinear and caused the models to fail to converge when included together in the regression equations.
7. We also examined the squared terms for royalty rates and franchise fees. However, we found no effects of for the survival value of these curvilinear terms.
Aldrich, H., and Fiol, M. 1994. “Fools Rush In? The Institutional Context of Industry Creation.” Academy of Management Review, 19(4): 645–670.
Bercovitz, J. 1997. “An Analysis of the Contract Provisions in Business-Format Franchise Agreements.” Unpublished Manuscript.
Bradach, J., and Eccles, R. 1989. “Price, Authority and Trust: From Ideal Types to Plural Forms.” Annual Review of Sociology, 15: 97–118.
Brickley, J., and Dark, F. 1987. “The Choice of Organizational Form: The Case of Franchising.” Journal of Financial Economics, 18: 401–420.
Brickley, J., Dark, F., and Weisbach, M. 1991. “The Economic Effects of Franchise Termination Laws.” Journal of Law and Economics, 34: 101–132.
Carney, M., and Gedajlovic, E. 1991. “Vertical Integration in Franchise Systems: Agency Theory and Resource Explanations.” Strategic Management Journal, 12: 607–629.
Carroll, G., and Delacroix, J. 1982. “Organizational Mortality in the Newspaper Industries of Argentina and Ireland: An Ecological Approach.” Administrative Science Quarterly, 27: 169–198.
Caves, R., and Murphy, W. 1976. “Franchising: Firms, Markets, and Intangible Assets.” Southern Economic Journal, 42: 572–586.
Dimaggio, P. and Powell, W. 1983. “The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields.” American Sociological Review, 48: 147–160.
Fama, E., and Jensen, M. 1983. “Agency Problems and Residual Claims.” Journal of Law and Economics, 26: 327–349.
Fladmoe-Lindquist, K., and Jacques, L. 1996. “Control Modes in International Service Organizations: The Propensity to Franchise.” Management Science, 41(7): 1238–1249.
Granovetter, M. 1985. “Economic Action and Social Structure: The Problem of Embeddedness.” American Journal of Sociology, 91: 481–510.
Hannan, M. and Freeman, J. 1984. “Structural Inertia and Organizational Change.” American Sociological Review, 49: 149–164.
Hybels, R. 1995. “On Legitimacy, Legitimation, and Organizations: A Critical Review and Integrative Theoretical Model.” In D. Moore (Ed.), Academy of Management Best Papers Proceedings, 241–245. Vancouver, B.C.: Academy of Management.
Ingram, P., and Baum, J. 1997. “Chain Affiliation and the Failure of Manhattan Hotels, 1898–1980.” Administrative Science Quarterly, 42: 68–102.
Kaufmann, D. 1992. “An Introduction to Franchising and Franchise Law.” Franchising: Business and Legal Issues. New York: New York State Bar Association.
Kennedy, P. 1992. A Guide to Econometric Methods (3rd edition). Cambridge, MA: MIT Press.
Lafontaine, F. 1992. “Agency Theory and Franchising: Some Empirical Results.” Rand Journal of Economics, 23: 263–283.
Lafontaine, F. 1995. “A Critical Appraisal of Data Sources on Franchising.” Journal of Marketing Channels, 4(1/2): 5–25.
Lafontaine, F., and Kaufmann, P. 1994. “The Evolution of Ownership Patterns in Franchise Systems.” Journal of Retailing, 70(2): 97–113.
Lafontaine, F., and Shaw, K. 1996. “The Dynamics of Franchise Contracting: Evidence from Panel Data.” National Bureau of Economic Research Working Paper, No. 5585.
Lafontaine, F., and Shaw, K. 1998. “Franchising Growth and Franchisor Entry and Exit in the U.S. Market: Myth and Reality.” Journal of Business Venturing, 13(2): 95–112.
Oliver, C. 1991. “Strategic Responses to Institutional Processes.” Academy of Management Review, 16(10): 145–179.
Podolny, J. 1993. “A Status-Based Model of Market Competition.” American Journal of Sociology, 98: 829–872.
Rao, H. 1994. “The Social Construction of Reputation: Certification Contests, Legitimation, and the Survival of Organizations in the American Automobile Industry, 1895–1912.” Strategic Management Journal, 15: 29–44.
Scott, F. 1995. “Franchising vs. Company Ownership as a Decision Variable of the Firm.” Review of Industrial Organization, 10: 69–81.
Shane, S. 1996. “Hybrid Organizational Arrangements and Their Implications for Firm Growth and Survival: A Study of New Franchisors.” Academy of Management Journal, 39(1): 216–234.
Shane, S. Forthcoming. “Making New Franchise Systems Work.” Strategic Management Journal.
Singh, J., and Lumsden, C. 1990. “Theory and Research in Organizational Ecology.” Annual Review of Sociology, 16: 161–195.
Singh, J., Tucker, D., and House, R. 1986. “Organizational Legitimacy and the Liability of Newness.” Administrative Science Quarterly, 31: 171–193.
Stinchcombe, A. 1965. “Social Structure and Organizations.” In J. March (Ed.), Handbook of Organizations, 153–193. Chicago: Rand McNally.
Wiewel, W., and Hunter, A. 1985. “The Interorganizational Network as a Resource: A Comparative Case Study on Organizational Genesis.” Administrative Science Quarterly, 30: 482–496.
Williams, D. Forthcoming. “Incomplete Contracting and Ex-Post Opportunism: Evidence from Franchise Contract Terminations.” Journal of Business Venturing.
Williamson, O. 1985. The Economic Institutions of Capitalism. New York.
© 1998 Babson College All Rights Reserved. Last Updated January 1999