Page 1
Empirical Evidence on Franchisee Exits in Mature U.S. Franchise Systems
Robert E. Stassen
University of Arkansas Department of Marketing & Logistics
302 Walton College of Business Fayetteville, AR 72701
+1 479 575-6155 (Phone) +1 479 575-8407 (Fax)
[email protected]
Marko Grünhagen
Eastern Illinois University School of Business 4002 Lumpkin Hall
Charleston, IL 61920 +1 217 581-6906 (Phone) +1 217 581-7244 (Fax) [email protected]
Page 2
Empirical Evidence on Franchisee Exits in Mature U.S. Franchise Systems
INTRODUCTION
Potential franchisees are attracted to franchising due to the reduced risks of investing in a
successful franchised system. As the International Franchise Association states as an advantage
to franchising is “a franchise increases your chances of business success because you are
associating with proven products and methods.” These advantages provide support for
governmental involvement in regulating the franchising sector, as well as supporting individual
investment through governmental grants and loans, based on the claims of superior survival
rates, or reduced risks. While our society has long accepted the general definition that proven
formats should, by definition, have lower failure rates, at this time very little has been reported
with respect to unit success within franchise systems, even for the most successful franchise
systems, where these rates should be the lowest.
Difficulty in evaluating franchise failure is in part related to the support provided by the
franchise system in facilitating transfers of locations between current franchisees and/or
repurchase, or temporary operation, as company-owned outlets. Specifically, the location may
be acceptable, but inexperience in operating the store may result in the transfer to of expectly
good location to another franchisee. When neither the unit can be transferred to another
franchisee, nor can conditions be prove to made acceptable for franchisor operation, the
franchisee and the locations are listed as franchise exits, in the annually submitted Uniform
Franchise Offer Circular, which is equivalent to a franchise failure.
This study examines these exit rates for franchisee locations for two of the U.S.’s largest
and best established franchise systems, linking these exit rates to the characteristics of the
Page 3
franchise system and to the competitive conditions in their geography. Since the systems are
well-established, successful franchise formats, these failure rates would have to viewed as those
which would establish the lower bounds of failure rates, as it would be inappropriate to attribute
any problems to the franchise concept itself, but (1) the capabilities of the franchisee exhibited
by the structure of the franchisee’s system, or (3) the conditions of competition faced either from
that within the brand or within the geographic market. The paper begins with a review of
literature pertaining to franchising and survival, contrasting those issues underlying the franchise
internal system’s structure and the characteristics of the competitive market. Following this,
descriptive statistics on are provided on the structure of unit ownership within the U.S.’s top two
franchise systems, with this data linked to the exit data to these structural classifications. Two
separate cross-sectional analyses are present which regress failure rates on the ownership
structure of the system and the competitive market structure an across the 51 U.S. states (and
District of Columbia) and across 437 of the U.S.’s Metropolitan Statistical Areas. Results show
that the proportion of single-unit franchisees is significantly associated with failures rates, but
controlling for this, substantial difference related to system size and the effect of coverage are
discussed.
FRANCHISING AND FRANCHISEE INCENTIVES
A franchise relationship is commonly considered a hybrid between autonomy and
dependence. Franchisees of various types exist, often distinguished by the size of their operation,
and the modalities of the contractual agreement with the franchisor (Kaufmann and Dant 1996).
They typically pay an entry fee as well as recurring royalties and advertising fees to the
franchisor. In return, franchise owners receive the right to use the trademark or even the entire
Page 4
business format as well as a host of services provided by the franchisor, often including staff
training, centralized booking services and hotlines, market analysis assistance, software support,
and the like.
Two primary arguments for the existence of a franchised business have been developed
in the academic literature, resource scarcity and agency theory. Encountering difficulties in
raising capital and developing managerial talent, i.e., “resource scarcity”, during the early stages
of a franchise system’s existence, franchisors recruit franchisees that will invest the money
needed and undertake a portion of the ensuing risk. The constraints of money, employee labor
and management are reduced considerably for franchisors if they can find franchisees that are
responsible for building and staffing units so that the franchisors can focus on the development
of the system and the brand (Carney and Gedajlovic 1991; Caves and Murphy 1976; Oxenfeldt
and Kelly 1968/69; Shane 1996). The predominance of multi-unit franchising has reinforced the
principle of separation of ownership into the franchisee becoming the operations and
development specialist and the franchisor becoming more focused on the upstream, or brand
management responsibilities. As an example, McDonald’s has steadily reduced the number of
its company-owned outlets from 2,100 in 2002 in the U.S. to 1,550 from over 2,100 in 2002,
while growing its U.S. number of franchised outlets from 11,228 to 12,477.
The consistent explanation with the shift to multi-unit franchising is provided by agency
theory (Eisenhardt 1989; Jensen and Meckling 1976) such that franchisees are (1) less apt to
shirk their responsibility because they have a stake in the business and (2) better suited to
monitor the retail end of the operations costs to provide superior profitability. Their incentives
are directly linked to their performance, which is typically not the case with managers of
businesses. When the business is owned by a franchisee, the owner’s livelihood is directly linked
Page 5
to the performance of the unit, so that the franchisor’s costly need for monitoring individual units
is greatly reduced (Fama and Jensen 1983; Norton 1988; Rubin 1978).
In a recent study, agency theory and resource scarcity have been integrated by
Castrogiovanni, Combs and Justis (2006), indicating that a paradigm shift from the resource
scarcity to the agency view might occur over time. While resource constraint concerns prevail
during the early stages of the franchise system’s life cycle, after an initial expansion, monitoring
and the related costs appear to move into the franchisor’s focus and agency motives start guiding
the continued use of franchising, rather than reverting units back to the franchisor.
From the franchisee’s perspective, the benefits of franchising in general lie in the proven
concept and system. The availability of franchisor support services as a strong motivator for
individuals to be attracted to franchising as a business concept has been noted elsewhere (Dant
1995; Dant and Peterson 1990). Aspiring franchisees appear to be attracted to the perceived
security that they associate with the franchisor backing them while also being dependent on their
success with the overall system. However, it has also been shown elsewhere (Grünhagen and
Dorsch 2003) that franchisees tend to realize that many of their initial expectations may not
come true over time, hence creating the potential for disappointment and conflict in the
relationship, potentially leading to premature exit from the system.
FRANCHISING AND SYSTEM FAILURES
The distinction between what constitutes franchise system failure and franchisee failure is
complex, yet significant (Holmberg and Morgan 2005). Frazer and Winzar (2005) suggest that it
may be difficult to determine whether a franchise system fails, as the company may cease to
exist, or franchisees might sell back their units to the company or to other franchisees. In other
Page 6
words, the failure of an entire system may well be linked to an exceedingly high turnover rate
among its franchised units. However, no consistent definition of franchise system failure has
been agreed upon in the literature. This ambiguity, paired with the plethora of causalities that
may contribute to the outright failure of entire franchise systems, may well be the reason that has
prompted researchers to develop alternative concepts. In a study conducted by Frazer (2001) of
Australian franchisors, two primary causes of “disruption” to franchise systems were examined,
lack of franchisor support and stage of the franchise life cycle. “Disruptions” to franchise
systems are defined by Frazer (2001) as legal disputes between franchisee and franchisor, and
franchised outlet conversions to some other form of ownership (whether company owned,
converted to another independent operation, or closed altogether). Reasons for system disruption
that have been identified include the size and age of the franchise system (Frazer 2001, Frazer
and Winzar 2005). As systems mature and expand, substantial disputes between franchisor and
franchisee may either cause the franchisor to take over franchised units, or may provide reason
for a franchisee to break away from the system and begin to operate as an independent business.
It has also been shown that franchisees are more likely to leave their system when the associated
start-up cost is relatively low (Frazer and Winzar 2005), i.e., in cases where there are few
requirements to invest in transaction specific assets. Adding to this emerging literature stream on
franchise system disruptions, Grünhagen, DiPietro, Stassen and Frazer (2008) found in a cross-
country comparison between the U.S. and Germany that system disruption was dependent on the
maturity of the market, and that the provision of services such as field visits and newsletters were
related to greater disruptions.
On the other hand, franchisee failure was systematically investigated in a recent study by
Michael and Combs (2008). Franchisee failure was shown to decline with prior industry
Page 7
experience, the requirement for franchisees to actively manage their outlets, exclusive territories,
and lower royalty rates (Michael and Combs 2008).
INTRABRAND FRANCHISE COMPETITION AND FAILURE RATES
Franchisors want an optimal number of units in a market to achieve adequate exposure
and convenience for their brand to maximize their royalities. A major challenge to the franchisor
is to maximizing royalties by increasing coverage without hurting the overall profitability of
their franchisees. Too much coverage, too much intrabrand competition, too little sales per unit,
should put additional pressure on the survival rates for franchisees, and particularly small, or
single-unit franchisees whose business is solely dependent on a single location.
One approach to minimize intrabrand competition is to have a market developed by
contracting with area developers (master franchisees) which become multi-unit franchisees
(Kaufmann and Kim 1995). The area developer agrees a minimum number of locations. The
franchisor bears the risks that the area’s units will be operated by a franchisee without a
demonstrated record of performance in the area while having the same agency problems as
would exist if the franchisor would operate the locations itself (Kaufmann and Dant 1996). With
a limited number of franchisees, or with the majority of locations with a single franchisee,
coverage may meet minimal requirement, but survival rates should be higher.
As opposed to area development, a predominant practice is the sequential assignment of
new franchise locations to existing franchisees, which is common in the limited-service food
industry. Kalnins and Lafontaine (2004) showed that across systems, contiguity was the most
significant factor related to the franchisee selected, such that clusters units emerge which
provides an area of market exclusivity for the franchisee. Separately, when survival rates are
examined, the local experience of multiunit operators is shown to be a highly significant
Page 8
determinant of success (Kalnins and Mayer 2004). For these reasons, multiunit franchisees are
likely to develop larger, contiguous trade areas for their units, minimizing the extent of
intrabrand competition possible, and increasing the difficulty of franchisors to maintain or
increase the coverage within metropolitan markets. Franchisees are satisfied with this approach,
up to the point where they may see the marginal costs of managing an additional outweigh the
benefits in system profitability. Problems, or conflict may emerge when a single-unit franchisee
refuses to develop and adjacent location, then the location would be offered to an adjacent
franchisee (Stassen and Mittelstaedt 1996). In this way, competing multiunit operators may
achieve greater coverage of markets for a single brand than might be possible through coverage
done wholly through single-unit operators, or through coverage by a single multiunit franchisee.
The result of the increased coverage is lower sales per store, and lower sales per the single-unit
franchisee.
RESEARCH PROPOSITIONS
Multi-unit ownership demonstrates a franchisee capabilities and potential for success in a
market. In some sense, the relationship between failure rate and the number of units operated by
a franchisee is near definitional if the franchise system is growing and better locations are given
to better franchisees. Conversely, those franchisees that remain single-unit operators are most
likely to experience failures. States and MSAs that have an above average proportion of single-
unit franchisees will have an above average proportion of franchise failures. Similarly, states
and MSAs with above average proportions in larger multi-unit systems will have lower, or below
average failure rates.
Page 9
The more concentrated the ownership of franchise outlets, the less likely a market will
become overstored, or saturated, and the less likely the market will experience above average
failure rates. Specifically, the more units owned by the largest franchisee in a market, or the
greater the proportion of total units within the largest franchisees, the lower the failure rates,
such that concentration ratios should be negatively related to failure rates. Similarly, states or
markets with below average coverage, have lower intrabrand competition, and would have lower
failure rates, such that coverage rates, specifically the number of units per person is positively
related to failure rates.
Franchisees have distinctive marks, but they are not immune from competition of similar
restaurant concepts. Limited service restaurant customers have shorter purchase cycle, place
more emphasis on convenience and location, and are faced with a broad array of choices. As the
number of competing restaurants grows, and if this results in lower per sales per restaurants in
states and MSAs, the results will lead to higher failure rates for franchised outlets. With regard
to measures of market competition, we would expect that in markets that had less inter-type
competition, there would be lower failure rates. As such, markets with higher sales per store or
experiencing lower growth rates would have lower failure rates.
FRANCHISE EXITS AND SYSTEM STRUCTURE
This study combines exit data reported in the 2007 Uniform Franchise Offer Circular (UFOC)
filings for McDonald’s and Subway, filed with the State of Illinois. The name of the franchisee,
the city, and phone numbers, were listed for 355 McDonald’s and 549 Subway locations. These
units were matched against listing for all franchisees for both chains obtained from a private list
service and assigned to a State and to a County so that it could be assigned to a Metropolitan
Statistical Area (MSA). There were 486 locations that could not be found on either list, 229 of
Page 10
these were matched to an existing franchisee as part of a multiunit system; the remaining 248
were classified as single-unit franchisees, 26 as McDonald’s and 222 as Subway. Three
locations could not be classified due to an incomplete or illegible city.
The assignment of failures, or exits, to franchisees is provided in Table 1. The
breakdown in the table shows the percentage by different sizes of multi-unit systems. While 26
(half) of the aforementioned exits could not be matched against the list for McDonalds, the 222
unmatched exits account for 82.5% for the single-unit franchisees exits for Subway, significantly
higher, indicating a more current listing for Subway. An additional noteworthy difference is the
significant higher exit rate for McDonald’s when comparing non-MSA location to MSA
locations.
Expectedly, exits fall heavily within single-unit franchisees for both systems, and it is
most noticeable for McDonald’s, but there are substantial differences between the two systems.
For McDonalds, single-unit franchisees account for roughly 5 percent of all franchised units, but
nearly 10% of the franchise exits. In contrast, single-unit franchisees account for roughly 24%
of all Subway units, with their failures being nearly half (269/546) of all failures, but the failure
rate within single-units is disproportionately low, at 6% (269/4,477). Further, while the larger,
10 or more multi-unit franchisee show failures, they are heavily concentrated within subset of
franchisees. For Subway, of the 214 10+ franchisees, 184 show no exits, with 19 having 1 exit,
but 11 with two or more, one of which with 10. In sharp contrast, McDonald’s 175 10-plus unit
franchisees, 51 show to have experienced a failed location, with all of these limited to a single
exit.
Page 11
ANALYSIS
Analysis of the exit rates was done with two cross-sectional approaches, across the 50
U.S. states and the District of Columbia (n=51), and across the Economic Census Metropolitan
and Micropolitan Statistical Areas, which span state lines and whose boundaries are formed by
counties. The 28,946 locations were analyzed separately by franchisee system, with each system
first aggregated by franchisee size, independent of state or MSA (producing Table 1), where the
unit size of the system was retained for classification of franchisee size for analysis within States
and MSA. Stated differently, if a Subway franchisee had 12 units in total, with 4 in Lincoln,
Nebraska, 4 in Omaha, Nebraska, and 4 in Council Bluffs, Iowa, those units would be counted as
locations operated by 10 or more-unit franchisees in the Iowa (4) and Nebraska (8) measures for
the state analysis and to the Lincoln (4) and the Omaha-Council Bluffs (8) MSA analysis. For
both the state and MSA analysis, a two-step aggregation procedure produced the structural
measures, first producing single franchisee size measures, and then aggregating into 51 state and
MSA observation files.
In addition to the size of franchisee measures, the largest franchisee and top four firm
concentration measures were created using the aggregation procedure. First, each states or
MSA’s largest franchisee was identified based on the number of units; then the second, third, and
fourth sized franchisee in each market were selected to create a measure of four firm
concentration for each state and MSA. The four firm concentration ratio was used in the state
analysis and due to wide range in size of MSAs, the largest franchisee was the more relevant
measure.
Table 2 provides the descriptive statistics for the measures included in the analysis. In all
cases exit rate was calculated as the number of exists divided by all locations available for
Page 12
analysis in either the state or the MSA. For all proportional measures (P single-unit, P 5+ unit,
P largest, P top 4, etc.), the total units ever assigned to a state or MSA were used for the
denominator. To assess the intensity of intrabrand competition within the states or MSAs, the
number of units per person was calculated using the Census estimates for 2007 for the states and
the summation of zip populations for all zip areas with establishments in each county, from
Census estimates. The U.S. Economic Census Reports for 2007 (for Hospitality and
Restaurants) provides the sales and number of establishments for the NAICS classification
722211, “Limited service restaurant” (LSR) was used for the states and 72221 “Limited-service
restaurant and refreshment places” was used for the MSAs due to missing data and would limit
analysis to 156 largest MSAs. Further, both the numbers of establishments and sales of each
classification are highly correlated within and across years and type, ranging from .957 to .995.
These census values were used to create the sales per establishment (Sales per LSR) and 2002-
2007 growth rate in establishments (2002-07 growth in LSR) measures.
In the state analysis, additional measures were included. To allow for an effect
attributable to being one of the 13 states requiring UFOC disclosures, a dummy variable
(1=UFOC state) was included. Second, as part of the aggregation of franchisee structure in the
state, the proportion of McDonald’s or Subway’s franchisees within MSAs, to examine the effect
of states franchising being in primarily urban markets.
RESULTS
Multiple regression analysis was conducted using the measures permitted from the
structural of ownership measures as well as other measures available from the Economic
Censuses, with those results in Tables 4 and 5. Those regression models in these tables were the
Page 13
results of testing competing models with fewer independent variables, with those reported having
the (1) highest variance explained with respect to adjusted R2, and (2) all regression coefficients
would retain the sign consistent with its bivariate correlation coefficients reported in Table 3.
Further, no variable would be retained in the analysis if it could not be shown to be highly
statistically significant in at least two regression analyses for one of the two franchise systems.
Tables 3 permits comparison of bivariate correlation coefficients between the two
franchisees, reading across, and the differences shown across state and MSA, reading top down.
No correlation among independent measures accounts for more than half of the variance in any
other (all are below .70), with the highest correlations occurring within the structural measures
from the MSA analysis. It should be noted that the magnitude of correlation must be higher in
the state analysis to be highly significant due to 8 times the observations available for analysis
with the MSAs. The majority of correlations are consistent in magnitude and significance when
making comparisons within franchise system and the relationships in the state analysis and the
MSA analysis. The proportion of units within states’ MSA has a significant effect on nearly
every measure in the study, but interestingly, has no effect on sales per LSR, suggesting some
equivalence across the states on this important characteristic related to failure.
Most consistent are the positive relationship between the proportion of single-unit
franchisees and failure rates and the negative relationship between the proportion establishment
in 5-9 unit franchisees and failure rate across states and MSAs. System coverage (units per
person) and several other measures show important differences on the effect of nearly every
measure and failure rate, particularly with regard to units per person, failure rates and the factors
affecting both.
Page 14
Table 4 provides the across-state results of regressing failure rates on franchisee
structures and market characteristics for the 51 observations. Due to the limited number of
observations, and high correlations shown with the independent measures and the proportion of
sales within MSA, these analyses kept the number of independent measures used in the analysis
to a minimum, as with one significant predictor, 24 observations and 5 independents provides a
negative adjusted R2 for the model with a lower proportion of units in MSAs.
For McDonald’s, the results show the predicted positive effects of proportion single-unit
and negative effects of the proportion of 5-9 unit on failure rates, with near identical results
shown for Subway. Noteworthy in the Table is the positive effect of concentration (P Top 4) on
failure rate for McDonald’s, in the combined 51 observation analysis (and no effect for in the
Subway analysis). Similarly, higher coverage (units per person), reduces failure rates (negative
coefficeints) for McDonalds, but has no effect for Subway. Sales per limited store restaurant
show mixed results, depending on model, but is generally positive, indicating that states with
higher sales per LSR have higher failure. No significant result is evident from the UFOC
dummy.
Table 5 provides the results of regressing failure rate on similar variables, but breaks the
analysis into subgroups of below <100,000; 100,000 to 250,000; and above 250,000. The
predictive ability of the models is substantial lower in the MSA analysis than in the states, with
the models for Subway substantially lower for Subway than for McDonalds in nearly every
specification. In terms of the structural measures, the proportion of single-unit franchisees again
is shown to be positively related to failure rates, except in the largest metro areas, with the
majority of restaurants and unit failures, the effect is not significant for either system, appearing
to exist only in the lower two-thirds of MSAs. Similarly inconsistent is the effect of the
Page 15
proportion of 5 or more unit systems, highly significant, and negative for the 100,000-250,000
market for McDonald’s.
In terms of competitive effects, the proportion of units within the largest franchisee (P
Largest) is shown to be negatively related to failure rate for McDonald’s, but not significant
factor for Subway. Conversely, coverage (units per person) has a negative effective in the mid
and lower sized market for McDonald’s, but is the best predictor of failure rates for Subway,
appearing to increase failure rates in all but the largest market. Also noteworthy is the negative
effect on failure rates for Sales per store, significant across the largest market for McDonalds,
and the positive effect on failure rates for Subway for growth rates for LSR in markets.
DISCUSSION
To date, this is the first study to attempt to relate system failure rates to the characteristics
of structure of ownership within the firm, or the characteristics of the market. In terms of failure
rates themselves, these are highly successful franchise systems, and should represent the lowest
expected failure rates for any mature franchise system, with results of “system exits” not
attributable to the characteristics of the franchise system.
Since the regression results show that in most specifications and for both systems,
instances of markets with an above average proportion of units in single-unit systems will have a
higher proportion of failures. Specifically, if you were a potential new single-unit franchisee for
McDonald’s or Subway, your probabilities of failure are approximately 10% and 6%,
respectively. Provided the franchise can demonstrate success and have an opportunity to open an
additional unit, these probabilities of new units failures drop significantly, given his experiences
experience. There’s no question that single-unit franchisees are associated with higher exit
Page 16
rates, but this may be a combined effect that the vast majority of new franchisees begin as single-
units, and by being new, face the more extreme tests of survival. If they choose to remain as a
single-unit operators, and not look to help the franchisor develop new locations, their
probabilities of survival make be tested even further, as the site they “passed on” may be offered
to a fellow franchisee, who now becomes a multi-unit competitor to the single-unit operator.
In some respects, there is more evidence for this type of intrabrand competition occurring
with Subway than with McDonalds, specifically, the state and MSA data show negative effects
on failure rate for coverage for McDonald’s, whereas the opposite is shown for Subway.
Similarly, MSAs with above average shares of units in the largest franchisee showed lower
failure rates for McDonald’s, and no effect for Subway, even though both systems exhibit similar
relationships between ownership structure and market coverage.
Page 17
RERERENCES
Castrogiovanni, G.J., J.G. Combs and R.T. Justis (2006). Shifting Imperatives: An Integrative
View of Resource Scarcity and Agency Reasons for Franchising. Entrepreneurship
Theory & Practice, 30(1), 23-40.
Caves, R. E. and W. F. Murphy (1976). Franchising: Firms, Markets, and Intangible Assets.
Southern Economic Journal, 42(April), 572-586.
Dant, R.P. (1995). Motivations for Franchising: Rhetoric versus Reality. International Small
Business Journal, 14(1), 10-32.
Dant, R.P. and A. Peterson (1990). Perceived Advantages of the Franchise Option from the
Franchisee Perspective: Empirical Insights from a Service Franchise. Journal of Small
Business Management, 28(July), 46-61.
Eisenhardt, K. (1989). Agency Theory: An Assessment and Review. Academy of Management
Review, 14(1), 57-74.
Fama, E. and M. Jensen (1983). Agency Problems and Residual Claims. Journal of Law and
Economics, 26, 327-349.
Frazer, L. (2001). Causes of Disruption to Franchise Operations. Journal of Business Research,
54(3), 227-234.
Frazer, L., and H. Winzar (2005). Exits and Expectations: Why Disappointed Franchisees Leave.
Journal of Business Research, 58(11), 1534-1542.
Grünhagen, M., DiPietro, R. B., Stassen, R.E., & Frazer, L. (2008). “The Effective Delivery of
Franchisor Services: A Comparison of U.S. and German Support Practices for
Franchisees”. Journal of Marketing Channels, 15(4), 315-335.
Page 18
Holmberg, S.A., and K.B. Morgan (2005). Entrepreneurial Franchise Ventures: US and
European Franchise Failure Strategic Perspectives and Empirical Investigation.
Proceedings of the International Society of Franchising Conference, London, UK.
Jensen, M.C. and W.H. Meckling (1976). Theory of the Firm: Managerial Behavior, Agency
Costs and Ownership Structure. Journal of Financial Economics, 3(4), 305-360.
Kalnins, Arturs and Francine Lafontaine (2004), “Multi-unit Ownership in Franchising:
Evidence from the Fast-Food Industry in Texas,” RAND Journal of Economics, 35 No. 4
(Winter) pp. 749-763.
Kalnins, Arturs and Kyle J. Mayer (2004), “Franchising, Ownership, and Experience: A Study
of Pizza Restaurant Survival,” Management Science, 50 No. 12 (December) pp. 1716-
1728.
Kaufmann, P.J. and R.P. Dant (1996). Multi-Unit Franchising: Growth and Management Issues.
Journal of Business Venturing, 11(5), 343-358.
Kaufmann, Patrick J. and Sang Hyeon Kim (1995), “Master Franchising and System Growth
Rates,” Journal of Marketing Channels, 4, Nos. 1-2, 49-64.
Michael, S.C. and J.G. Combs (2007). Entrepreneurial Failure: The Case of Franchisees. Journal
of Small Business Management (forthcoming).
Norton, S.W. (1988). Franchising, Brand Name Capital, and the Entrepreneurial Capacity
Problem. Strategic Management Journal, 9(Summer), 105-114.
Oxenfeldt, A.R. and A.O. Kelly (1968/69). Will Successful Franchise Systems Ultimately
Become Wholly-Owned Chains? Journal of Retailing, 44(Winter), 69-83.
Rubin, P. (April, 1978). The Theory of the Firm and the Structure of the Franchise Contract.
Journal of Law and Economics, 21, 223-233.
Page 19
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.
Stassen, Robert E. and Robert A. Mittelstaedt (1995), “Territory Encroachment in Maturing
Franchise Systems,” Journal of Marketing Channels, 4, Nos. 1 and 2, 27-48.
Page 20
Table 1 Subway and McDonald’s Exits by Franchise Multi-Unit Ownership Type
McDonald’s Franchisees
Subway Franchisees
Number Exits Units Rate Number Exits Units Rate
Single unit 527 52 527 9.9%
4,477 269 4,477 6.0%
2 unit 555 50 1,110 4.5%
1,567 67 3,134 2.1%
3-4 units 689 92 2,339 3.9%
1,061 84 3,568 2.4%
5-9 units 580 110 3,627 3.0%
583 70 3,701 1.9%
10 or more 175 51 2,613 2.0%
214 56 3,850 1.5%
2,526 355 10,216 3.5%
7,902 546 18,730 2.9%
MSA
322 9,861 3.3%
513 17,379 3.0%
Non MSA
33 355 9.3%
33 1,351 2.4%
Page 21
Table 2 Descriptive Statistics for Separate Cross-Sectional Analyses
Cross Sectional Analysis of States McDonald's Minimum States Maximum States Mean Std. D.
Exit rate .00 3 states .09 Alaska .037 .019
P single-unit franchisees
.00 7 states .11 Pennsylvania .045 .029
P 5-9 unit franchisees .00 Alaska .67 Nebraska .366 .129
P Top 4 franchisees .06 California 1.00 Vermont .351 .207
Units/10,000 persons .16 Alaska .48 Montana .352 .066
P units in MSAs .65 Montana 1.00 9 states .905 .095
Subway
Exit rate .00 14 states .19 Hawaii .029 .035
P single-unit franchisees
.03 Kansas .55 Hawaii .236 .101
P 5-9 unit franchisees .00 Hawaii .50 Alabama .246 .105
P Top 4 franchisees .03 California .88 Alaska .231 .188
Units/10,000 persons .35 New Jersey .89 Nebraska .656 .115
P units in MSAs .59 Vermont 1.00 9 states .898 .101
Sales per limited store restaurants
$479 Rhode Island $884 New Mexico $700 $99
Cross Sectional Analysis of Metropolitan Statistical Areas
McDonald's, N=417 Minimum MSAs Maximum MSAs Mean Std. D. Exit rate .00 268 MSAs 1.00 2 MSAs .042 .104
P Single-unit franchisees
.00 264 MSAs 1.00 3 MSAs .049 .116
P 5 + unit franchisees .00 68 MSAs 1.00 101 MSAs .617 .345
P Largest franchisees .00 2 MSAs 1.00 96 MSAs .589 .299
Units/10,000 persons .04 Port St. Lucie, FL 1.78 Gallup, NM .525 .218
Subway, N=437
Exit rate .00 370 MSAs .750 LaGrange, GA .019 .076
P Single-unit franchisees
.00 81 MSAs .750 Rome, GA .197 .158
P 5 + unit franchisees .00 47 MSAs 1.00 20 MSAs .454 .301
P Largest franchisees .02 New York-Wayne-White Plains, NY-NJ
1.00 Danville, KY .369 .241
Units/10,000 persons .31 St. Joseph, MO 2.09 Branson, MO .917 .270
Sales per limited store restaurants
$284 Ocean City, NJ $1,019 Midland, TX 700 113
2002-07 growth in LSR -12.2% Ardmore, OK 57.6% Kingsport-Bristol, TN-VA
17.4% 12.0%
Page 22
Table 3 Inter-item Correlation Coefficients of` Franchisee Ownership and Market Characteristics
McDonald’s, State Analysis, n=51 Subway, State Analysis, n=51
Failure rate
Single-units
5 to9 units
Top 4 fr’ees. Coverage
Sales per LSR
Failure rate
Single-units
5 to9 units
Top 4 fr’ees. Coverage
Sales per LSR
Failure rate 1
1
P single-units .21 1
.51 3 1
P 5-9 units -.42 3 .03 1
-.40 3 -.45 3 1
P Top 4 .17 -.34 2 -.26 1 1
-.27 1 -.46 3 -.10 1
Units/10,000 persons
-.21 .06 .01 -.01 1
-.07 -.36 3 .42 3 -.06 1
Sales per LSR .06 -.22 .13 -.30 2 .31 2 1
-.04 -.43 3 .18 -.06 .38 3 1
P MSA units -.14 -.32 2 -.29 2 -.17 -.35 2 -.07 1
.26 1 .32 2 -.46 3 -.26 1 -.32 .04 1
McDonald’s, MSA Analysis, n=424
Subway, MSA Analysis, n=437
Failure rate
P single-units
P 5 + units
Largest fr’ee Coverage
Sales per LSR
Failure rate
Single-units
P 5 + units
Largest fr’ee Coverage
Sales per LSR
Failure rate 1
1
P single-units .43 3 1
.20 3 1
P 5+ units -.12 2 -.25 3 1
-.14 3 -.64 3 1
P Largest -.22 3 -.21 3 .25 3 1
-.13 3 -.55 3 .59 3 1
Units/10,000 persons
-.11 2 -.21 3 .20 3 .16 3 1
.17 3 -.15 3 .20 3 .12 3 1
Sales per LSR -.05 -.05 .07 .04 .08 1
-.01 -.23 3 .21 3 .13 2 .00 1
2002-07 growth in LSR
-.05 -.03 .04 -.04 -.04 -.07 1
.09 .00 .01 -.04 -.04 -.07 1
1 p<.10; 2 p<.05; 3 p<.01
Page 23
Table 4 Failed Establishments as a Proportion of Total in State Regressed on State Characteristics
Standardized Regression Coefficients and t-Statistics
All States
Lower pMSA (<.92)
Higher pMSA >.92 McDonalds’ β t β t β t
(Constant) .79
1.02
.47 P single-unit
franchisees .43 3.12 3
.15 .64
.49 2.31 2
P 5 -9 unit franchisees -.39 -3.28 3
-.33 -1.46
-.46 -2.86 3
P top 4 franchisees .31 2.13 2
.20 .74
.37 1.67
Units/10,000 persons -.36 -2.90 3
-.39 -1.75 1
-.30 -2.00 1
Sales per LSR .40 2.77 3
.20 .70
.42 2.39 2
UFOC -.08 -.64
N 51 24 27
R2
.42
.20
.61
Adj. R2
.34
-.02
.51
Subway
Lower pMSA (<.92)
Higher pMSA >.92 (Constant) -.91
-1.16
-.67
P Single-unit franchisees
.48 2.43 2
.20 .92
.59 2.29 2
P 5 -9 unit franchisees -.29 -1.86 1
-.18 -.88
-.34 -1.67
P top 4 franchisees -.06 -.37
-.10 -.39
.12 .59
Units/10,000 persons .16 1.13
-.22 -1.07
.16 .97
Sales per LSR .15 1.03
.72 2.74 2
.06 .32
UFOC -.01 -.08
N 51 24 27 R2
.36
.48
.53
Adj. R2
.27
.33
.42
1 p<.10; 2 p<.05; 3 p<.01
Page 24
Table 5 Failed Establishments as a Proportion of Total in MSA Regressed on MSA Characteristics
Standardized Regression Coefficients and t-Statistics
All MSAs <100,000 100,000-250,000 >250,000 β t β t β t β t
McDonald’s (Constant)
2.27 2
4.01 3
2.25 2
4.15 3
P Single-unit franchisees
.40 8.73 3 .32 4.06 3 .41 4.98 3 -.12 -1.36
P 5+ unit franchisees
.02 .35 .04 .59 .26 2.78 3 -.17 -1.71 1
P Largest -.14 -3.01 3 -.39 -4.64 3 -.41 -4.47 3 .04 .42
Units/10,000 persons
.00 -.07 -.13 -1.79 1 -.22 -2.53 2 .02 .21
Sales per LSR -.03 -.60 -.03 -.45 .04 .54 -.19 -2.38 2
2002-07 growth in LSR
-.05 -1.17 .03 .44 -.08 -1.11 -.01 -.06
N
424
133
130
161
R2 .21 .36 .35 .06
Adj. R2
.20
.33
.32
.02
Subway
(Constant) -2.74 3
-2.06 2
-2.16 2
.32
P Single-unit franchisees
.19 3.08 3 .13 1.29 .25 2.46 2 .14 1.03
P 5+ unit franchisees
-.06 -.88 -.12 -1.08 -.04 -.29 -.06 -.41
P Largest -.02 -.40 .02 .18 .00 .03 .05 .49
Units/10,000 persons
.22 4.73 3 .25 2.86 3 .27 3.01 3 .12 1.42
Sales per LSR .05 1.09 .12 1.49 .05 .57 -.13 -1.50
2002-07 growth in LSR
.10 2.15 2 .06 .68 .13 1.53 .16 2.04 2
N
437
142
133
162
R2 .10 .11 .13 .08
Adj. R2
.09
.08
.10
.05
1 p<.10; 2 p<.05; 3 p<.01