Empirical Evidence on Franchisee Exits in Mature …emnet.univie.ac.at/uploads/media/Stassen_Grünhagen_05.pdfEmpirical Evidence on Franchisee Exits in Mature U.S. Franchise Systems

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]

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

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

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

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

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

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

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.

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

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.

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

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

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.

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

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

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.

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.

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.

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.

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%

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%

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

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

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