Assessing the Correlates of Performance in Franchised Channels of Distribution:
A Cointegration Analysis
Rajiv P. Dant
Associate Professor of Marketing
College of Business Administration
University of South Florida
4202 E. Fowler Avenue, BSN 3403
Tampa, FL 33620, USA
Voice: 813.974.6227
Facsimile: 813.974.6175
Email:
Manish Kacker
Assistant Professor of Marketing
Tulane University - Marketing Area
7 McAlister Drive
New Orleans , LA 70118, USA
Voice: 504.865.5494
Facsimile: 504.865.6751
Email:
Anne T. Coughlan
Associate Professor of Marketing
Kellogg School of Management
Northwestern University
2001 Sheridan Road
Evanston, IL 60208-2008, USA
Voice: 847.491.2719
Facsimile: 847.491.2498
Email:
Jamie Emerson
Assistant Professor of Economics
School of Business
Clarkson University
Potsdam, NY 13699, USA
Voice: 315.268.6457
Facsimile: 315.268.3810
Email:
Submitted to:
EMNET 2005: International Conference on Economics & Management of Networks
September 15-17, Budapest, Hungary
Assessing the Correlates of Performance in Franchised Channels of Distribution:
A Cointegration Analysis
Rajiv P. Dant, University of South Florida, USA
Manish Kacker, Tulane University, USA
Anne T. Coughlan, Northwestern University, USA
Jamie Emerson, Clarkson University, USA
ABSTRACT
Not much is known about the primary drivers of performance in franchising systems. With some notable exceptions, much of the franchising literature on performance related issues has focused on either contrasting failure rates of independent small businesses and entrepreneurs with those of franchises and/or system survival issues. Systematic investigations of the drivers of franchising performance are virtually non-existent. The existing literature base on franchising performance displays at least three other characteristic patterns. First, most studies have restricted themselves to a single sector, usually, the fast food restaurant industry, since it is often perceived and portrayed as the archetypical franchise system. Second, existing investigations have tended to focus on a single measure of performance. Finally, with the exception of survival articles, empirical studies have typically confined themselves to cross-sectional examination of the evidence. In other words, we know very little about what fosters long term performance.
Our investigation of the correlates of performance, then, contributes to the extant literature in three specific ways. Foremost, we attempt a systematic assessment of the relative effects of a series of firm decision variables on performance. Specifically, we evaluate the impact of four categories of drivers of performance: (1) strategic decisions regarding business goals, (2) strategic decisions involving governance structure and vertical restraints, (3) strategic decisions related to the marketing strategy in the nascent stages of the system development, and (4) strategic decisions related to the ongoing marketing strategy. Besides three covariates, a total of eleven hypotheses focused on drivers of performance are investigated. Second, we utilize three different operationalizations of our dependent variable, performance, in our investigation. Third and finally, we estimate our empirical models using nine years of longitudinal panel data aimed at deciphering the effects associated with our set of predictor variables using cointegration analysis, an exciting, new approach to modeling equilibrium or long term relationships between economic variables in panel data.
The results show that seven out of eleven hypotheses were supported by the data using the system size operationalization of performance.
Assessing the Correlates of Performance in Franchised Channels of Distribution:
A Cointegration Analysis
INTRODUCTION
Franchising has been cited as one of the fastest growing modes of retailing in the U.S. as well as the world.Although its growing significance to the economy has attracted research attention to it from diverse disciplines including management, law, economics, marketing, and finance, surprisingly fundamental knowledge gaps continue to characterize our current understanding of performance in business format franchising. Although extant research has looked at performance in franchise systems from multiple perspectives, a definitive understanding of such a multifaceted construct has remained elusive and necessitated calls for further investigation. This paper seeks to address some of the knowledge gaps related to the construct of performance within the business format franchising sector.
The earliest studies of performance in franchise systems operationalized performance as an attitudinal construct, for which self-reported measures were collected from franchisors and/or franchisees. Hunt and Nevin (1974) and Lewis and Lambert (1991) focused on ‘franchisee satisfaction’. Other researchers (e.g., Guiltinan, Rejab and Rodgers 1980; Anand and Stern 1985) have examined elements of franchisee cooperativeness. While these studies provide valuable insights into factors underlying harmonious and productive franchise relationships, they have drawn criticism for their use of attitudinal, self-reported measures of performance.
More recent studies have chosen to focus on archival data-based measures of performance in franchise systems. However, some of these studies are constrained by the industry-specific nature of their performance measures that limit comparisons across industries – for example, performance measures like ‘Revenues per room’ in the lodging industry (Kalnins 2004). Other studies use standardized financial metrics that can be compared across industries but involve samples limited to publicly traded companies, e.g., chain wide revenue growth (Sorenson and Sorenson 2001), return on assets and market-to-book value (Combs and Ketchen, Jr. 1999), shareholder returns (Leleux, Spinelli and Birley 2003), return on equity (Alon, Drtina and Gilbert 2004) economic value added (EVA) and market value added (MVA) (Aliouche and Schlentrich 2005). One limitation of these studies is that theirsamples of publicly traded franchisors tend to be a homogenous subsection of the franchising universe,and are arguably very different from the rest of the franchising world.
Contrary to the conventional wisdom propagated in the popular press, it is not easy for new franchisors and franchisees to survive. Failure rates for these organizations are often similar (and at times exceed) the failure rates for corresponding independent businesses in the U.S. (Castrogiovanni Justis and Julian 1993, Bates 1995, Shane 1996, Lafontaine and Shaw 1998, and Holmberg and Morgan 2003) as well as in the U.K. (Stanworth et al. 1998). These trends, coupled with the limitations associated with the previously mentioned attitudinal and archival measures of franchise system performance, have made the related constructs of failure and survival the focus of much of the extant research on franchise system performance. Studies such as those mentioned above as well as others (e.g., Bates 1998, Shane 1998a, Shane 1998b, Shane and Foo 1999, Shane 2001) have attempted to identify factors that affect the survival of new franchisors and/or franchisees.
Historically, a research agenda focused on either contrasting failure rates of independent small businesses and entrepreneurs with those of franchises and/or system survival issueswas probably inevitable as this relatively novel approach to business sought to legitimize itself as a viable economic activity. The repeated comparisons with independent small businesses also reflect a common practice within franchising to position itself to downstream would-be franchisees as an entrepreneurial activity.[1] Hence, the franchising sector has consistently felt the need to showcase its superiority to the alternative of independent entrepreneurial enterprise, and the available literature on franchising performance reflects this preoccupation.
Performance, however, is more than survival. In our view, performance is also about growth. There are a number of ways to measure growth in franchise systems. However, financial measures (e.g., sales, assets) or employment numbers are not as robust measures of franchise system growth as the number of outlets in the system (Martin and Justis 1993). Blair and Lafontaine (2005) agree that growth in the number of units in business format franchise systems is a suitable proxy for the growth in business format franchising, affirming Sen’s (1998) finding of a significant correlation between system outlet growth and system dollar sales growth.
Growth in franchise systems is not easily attained. Stanworth (1996) and Perrigot (2004) find that less than 50 percent of franchisors that survived continued to grow at healthy rates in the UK and France respectively. In the U.S., Blair and Lafontaine (2005 p. 23) find that, contrary to popular belief, growth in franchise systems has been modest and similar to that for the economy as a whole. Moreover, they find considerable variation in growth across franchise systems: in 2001, approximately 45% of franchisors operated systems with less than 50 units, and approximately 89% of franchisors operated systems with 500 or fewer units (Blair and Lafontaine 2005, p. 48). The findings from these studies raise the following questions: Why is it that only a small percentage of franchise systems grow to be large systems? What are the factors that influence the growth of these systems? What are the decisions that a franchisor can take to enhance this central aspect of franchise system performance?
Given the limitations associated with other measures of franchise performance and the elusiveness of system growth and success, it is not surprising that few extant studies have attempted to identify and understand drivers of franchise system growth and size. Shane (1996) analyzed data for 138 franchise systems and found that a greater emphasis on franchising as an organizational form resulted in faster system growth. In a study that covered 109 franchisors in the restaurant sector, Sen (1998) examined the impact of increased franchising proportion and system size on the change in number of system outlets between 1986 and 1990. He, too, found a positive association between increases in reliance on franchising and system growth. Castrogiovanni and Justis (2002) investigated the effects of some other strategic and contextual factors on a similar dependent variable for a sample of 246 franchise systems. Using a non-parametric approach characterized by the rank transformation of continuous independent variables prior to their inclusion in multiple regression analyses, they found two strategic factors (‘franchise start up costs’ and ‘franchisor growth orientation’) to have a significant positive impact on the percentage change in total units in a franchise network over a five period from 1993 to 1998. Significant negative effects were found for two contextual factors (‘industry growth’ and ‘franchisor age’).
Our investigation of the correlates of performance, then, contributes to the extant literature in three specific ways. First, we draw on agency theory, transaction cost analysis, resource scarcity and resource-based theories to systematically assess the relative effects of a comprehensive series of firm-level decision variables on performance. Specifically, we evaluate the impact of three categories of drivers of performance: (1) strategic decisions regarding firm goals (2) strategic decisions related to the marketing mix variables and (3) strategic decisions involving structure and governance of the franchise systems. Second, the empirical models are evaluated using nine years of longitudinal data from multiple industries. Third and finally, we use cointegration analysis to estimate our empirical models in order to account for the lags associated with our set of predictor variables.
HYPOTHESES
The efficient functioning of a franchise system can be hindered by the presence of franchisee free riding or ex-post franchisor moral hazard. Both of these disruptions can be curtailed through the presence of higher royalty rates in franchise contracts. These royalties provide franchisors with compensation that offsets costs associated with monitoring franchisees and enforcing contractual provisions to prevent franchisee free riding (Brickley and Dark 1987). Knott (2001) finds that the involvement of the franchisor as a ‘hierarchical manager’ is necessary for the active enforcement of routine and the introduction of innovation in franchised units in the system. Furthermore, higher royalty rates give the franchisor a bigger stake in the performance of the franchise system and act to reduce the likelihood of ex-post franchisor opportunism (Lal 1990). Gallini and Lutz (1992) view the presence of a high royalty rate to be a signal of confidence about the product demand for the concept from the franchisor to prospective franchises. The alignment of incentives, revelation of favorable private information about product demandand the presence of chain-wide efficiencies makes participation more attractive to franchisees, thereby facilitating system growth. Therefore, we posit:
H1: Royalty rate is positively related to franchise system growth.
Carney and Gedajlovic (1991) find that franchisors vary in terms of the speed with which they want to grow. Castrogiovanni and Justis (2002) found a positive relation between a franchisor’s growth orientation and the actual growth of its network. It seems logical, then that franchisors that seek to grow at faster rates will actually do so. Therefore:
H2:The aggressiveness of franchisor growth goals is positively related to franchise system growth.
Lafontaine and Shaw (1998) found the primary driver of franchisor survival to be the number of years that the franchisor has been in business before starting to franchise. The effort associated with simultaneously developing and refining a business concept as well as undertaking the recruitment of franchisees and the establishment of a franchise system can be overwhelming for franchisors. When a franchisor spends a greater amount of time developing the business concept before commencing franchising, it is often able to come up with a better quality franchise concept that rapidly engenders superior brand reputation and is therefore more attractive to franchisees. In addition, the franchisor is able to fully focus on the logistics of establishing a franchise system once it decides to franchise. Therefore:
H3:Concept development time is positively related to franchise system growth.
Consumer advertising is important for creating brand awareness. Higher levels of awareness among end-users can create enhanced demand for a larger network of units. High brand equity becomes a resource valued by prospective franchisees (Combs and Ketchen, Jr. 1999). In addition, this brand equity creates safeguards against ex-post opportunistic behavior by the franchisor. These two factors make franchise systems with high levels of consumer advertising more attractive to prospective franchisees. Therefore:
H4:The level of consumer advertising is positively related to franchise system growth.
Extant research on franchise system growth (e.g., Shane 1996, Sen 1999) has found a positive relation between the extent of franchising used by a franchisor and franchise system growth. This result is consistent with the resource scarcity view that franchising allows a franchisor to overcome capital, informational and managerial constraints and grow rapidly (Norton 1988). However, the signaling-based agency theory literature (Gallini and Lutz 1992) suggests the opposite: a franchisor is expected to signal favorable private information about product demand to prospective franchisees through a higher proportion of company-owned outlets. Therefore, we have competing hypotheses:
H5a:The proportion of franchised outlets is positively related to franchise system growth.
and alternatively,
H5b:The proportion of company-owned outlets is positively related to franchise system growth.
In recent years, franchisors have incorporated a number of growth initiatives into their adoption of franchising. These include the use of multi-unit franchising (in the form of area development agreements, sub-franchising or territorial expansion) and allowing for conversion. Multi-unit franchising promotes system growth by enabling franchisors to overcome resource scarcity and adverse selection problems (Kaufmann and Dant 1996, Norton 1988). In addition, as ascertained by Kalnins and Lafontaine (1996), it allows franchisees to internalize externalities and reduce spillover effects. Finally, Bercovitz (2002) notes that multi-unit franchising enhances the downstream rent potential for franchisees, thereby creating the front-end of self-enforcing agreements (Klein 1980). The subsequent alignment of incentives, lowering of agency costs and reduction of the likelihood of the use of contractually-specified disciplinary devices enhances the attractiveness of the franchise system to prospective franchisees.
Conversion franchising entails the recruitment of franchisees from other chains and franchise systems. Franchisors who are successful in using conversion franchising acquire experienced franchisees that ‘hit the ground running’ and facilitate system growth.
H6:The number of franchisor growth initiatives is positively related to franchise system growth.
Franchisors perform a number of tasks that directly and indirectly support franchisees, facilitating their success and the consequent growth of the franchise system. These include the exercise of rigor in selecting new franchisees, the provision initial assistance to these franchisees and the provision of ongoing services to all franchisees. These activities enhance brand equity for the system and mitigate adverse selection, thereby creating incentives for high quality franchisees to join and grow the system. Therefore:
H7:The extent of assistance provided by the franchisor to new franchisees is positively related to franchise system growth.
H8:The extent of rigor used by the franchisor to select new franchisees ispositively related to franchise system growth.
H9:The extent of ongoing services provided by the franchisor to all franchisees is positively related to franchise system growth.
Some franchisors allow for passive ownership on the part of franchisees. While this expands the pool of prospective franchisees, it also creates an additional layer of hierarchy and agency costs within the system. The high-powered incentives associated with franchising are diminished. Shirking at outlets franchised by passive owners is likely to be high, necessitating increased monitoring effort on the part of the franchisor. Therefore:
H10:The extent of use of passive ownership is negatively related to franchise system growth.
In addition to these eleven hypotheses, we also investigate the effects of three covariates on the outcome measure of system growth in an exploratory mode. Two of these covariates are commonly used in most inter-organizational for control purposes, namely, age and size of the organization in question since they are expected to impact various theoretically vested relationships in systemic fashion. The third covariate, proportion of units in the U.S., is unique to the context at hand in that franchising has its genesis in the U.S., and therefore its domestic proportion may have some systemic effects on the system growth.
DATA CHARACTERISTICS
We tested our 11 hypotheses using secondary data drawn from Bond’s Franchise Guides from 1985 to 2004 (i.e., sixteen years of data as there were no Guides for years 1986, 1987, 1990 and 2000). However, in a tradeoff involving the number of years of data and the list-wise sample size across those years for the variables of interest, we finally settled on the years 1995 through 2004 (i.e., nine years) with annual sample size of N=76.Consequently, the cointegration analysis, the principal inferential statistical technique utilized, is based on panel data with N = 684 (9 years x 76) cases.