Ownership, Agency, and Wages: An Examination of Franchising in the Fast Food Industry
Author(s): Alan B. Krueger
Source: The Quarterly Journal of Economics, Vol. 106, No. 1 (Feb., 1991), pp. 75-101
Published by: Oxford University Press
Stable URL: http://www.jstor.org/stable/2937907
Accessed: 06-04-2018 11:06 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
Oxford University Press is collaborating with JSTOR to digitize, preserve and extend access
to The Quarterly Journal of Economics
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
OWNERSHIP, AGENCY, AND WAGES: AN EXAMINATION
OF FRANCHISING IN THE FAST FOOD INDUSTRY*
ALAN B. KRUEGER
This paper estimates the difference in compensation between company-owned
and franchisee-owned fast food restaurants. The contrast is of interest because
contractual arrangements give managers of company-owned outlets less of an
incentive to monitor and supervise employees. Estimates based on two data sets
suggest that employee compensation is slightly greater at company-owned outlets
than at franchisee-owned outlets. The earnings gap is 9 percent for assistant and
shift managers and 2 percent for full-time crew workers. Furthermore, the
tenure-earnings profile is steeper at company-owned restaurants. These findings
suggest that monitoring difficulties influence the timing and generosity of
compensation.
A topic of controversy among economists in recent years has
been whether firms adjust the level and timing of compensation to
solve incentive problems. Becker and Stigler [1974], Lazear [1981],
and others have argued that firms initially pay wages below the
workers' alternative wage and later pay wages above the alternative
wage to discourage shirking when monitoring is imperfect.
Such a delayed-payment/bonding contract is efficient because it
does not alter the present value of compensation from the firstbest,
full-information level. If legal or other constraints (e.g., the
minimum wage) restrict up-front bonds, however, a related literature
predicts that firms will pay efficiency wages.' An efficiency
wage payment differs from a delayed-payment/bonding contract in
that firms raise the present value of compensation to increase the
cost of job loss and thereby discourage shirking, although compensation
may also be backloaded in the efficiency wage model.
*I thank David Bloom, Charlie Brown, Richard Caves, Richard Freeman, Bob
Gibbons, Zvi Griliches, Larry Katz, John Pencavel, Andrei Shleifer, Bob Smith,
Larry Summers, two referees, and seminar participants at the NBER, Harvard,
MIT, Yale, the University of Maryland, the Universities of California (Berkeley, Los
Angeles, and San Diego), Stanford, Princeton, the University of Rochester, and the
University of Pennsylvania for helpful comments. I am also grateful to the National
Institute for Work and Learning and the Bureau of National Affairs for providing
me with their data sets.
1. Shapiro and Stiglitz [1984] and Bulow and Summers [1986] present
efficiency wage models assuming that delayed-payment/bonding contracts are ruled
out. Eaton and White [1982] and Dickens, Katz, Lang, and Summers [1989] discuss
efficiency wages when there are constraints on delayed-payment/bonding contracts.
If there are constraints on up-front bonds, the optimal contract will backload
compensation to some extent, and pay a higher wage than the alternative wage (see
also Akerlof and Katz [1989]).
v 1991 by the President and Fellows of Harvard College and the Massachusetts Institute of
Technology.
The Quarterly Journal of Economics, February 1991
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
76 QUARTERLY JOURNAL OF ECONOMICS
Whether firms overcome monitoring deficiencies by having workers
post implicit bonds or by paying efficiency wages-or whether
imperfect monitoring and shirking are indeed concerns of many
firms-are questions that can only be resolved empirically.
Unfortunately, little empirical work exists exploring how firms
respond to imperfect monitoring and agency problems.2 This paper
examines the labor market in the fast food industry to estimate the
effect of agency problems on the structure of compensation. The
institutional features of the fast food industry provide unique
conditions for studying wage determination. The fast food industry
is competitive, homogeneous, and nonunion. But most importantly,
the fact that some restaurant outlets are owned and
operated by the parent company, while others are owned and
operated by individual franchisees generates variability in organizational
structure that allows for a test of theories of wage
determination.
It is argued that existing contractual arrangements give
managers of company-owned restaurants different incentives from
franchisees who typically own and manage their own restaurants.
An owner-manager of a franchise has a strong incentive to expend
effort supervising and monitoring his workers because he receives
the residual profit generated by the enterprise; whereas a manager
of a company-owned establishment is usually not paid a share of
the establishment's profit, and his actions are not perfectly observed
by his principal, the parent company. As Caves and Murphy
[1976, p. 575] and others have noted, the latter contractual
arrangement poses a classic principal-agent problem that is likely
to reduce the level of supervision and monitoring in companyowned
units vis-a-vis franchised units. Some casual evidence is
presented in the next section supporting the maintained assumption
that principal-agent problems create monitoring and supervision
difficulties in company-owned outlets.
Under the assumption that monitoring is less rigorous at
company-owned outlets than at franchisee-owned outlets, the
delayed-compensation model would predict that the tenureearnings
profile of workers is steeper at company-owned outlets,
but total discounted compensation is equal under either form of
ownership. The efficiency wage model, on the other hand, would
predict that both the slope of the tenure-earnings profile and the
present value of compensation are greater at company-owned
2. Exceptions are Leonard [1987], Cappelli and Chauvin [1988], Lazear and
Moore [1984], Hutchens [1986], and Groshen and Krueger [1990].
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
OWNERSHIP, AGENCY, AND WAGES 77
outlets. Finally, the standard neoclassical model would predict no
difference in the timing or level of compensation at companyowned
and franchised outlets because employee effort is assumed
to be exogenously determined. Two newly available cross-sectional
data sets are used to test these predictions.
The main empirical finding is that low-level managers earn 9
percent higher wages at company-owned outlets than at franchised
outlets, whereas crew workers earn 1 to 2 percent more at
company-owned outlets. Moreover, the earnings differences result
almost entirely from comparatively steeper tenure-earnings profiles
at company-owned outlets. In summary, it seems that starting
pay is about the same at company-owned and franchised restaurants,
but pay grows more rapidly at company-owned restaurants.
These findings support a conclusion that-especially for low-level
managers-a combination of high wages and delayed compensation
are used to elicit effort from employees in establishments
where monitoring is more difficult.
An alternative interpretation of these results is that they stem
from another agency problem, namely, the lack of incentive for
company-owned outlets to minimize costs. If managers receive
added utility by paying high wages (e.g., through reduced grievances),
one would expect to find higher wage rates at companyowned
outlets, since there is a tenuous link between a hired
manager's compensation and the profitability of his restaurant. In
other words, company-outlet managers do not bear the full cost of
sharing profits with workers. However, the fact that wages for
employees of company-owned outlets are established by district
managers rather than by outlet managers may reduce the plausibility
of the "expense-preference" model in this context.
The remainder of the paper is organized as follows. Section I
discusses the institutional features of the fast food industry,
emphasizing explanations for franchising. Section II discusses
implications of implicit bonds and efficiency wage models for wages
in the fast food industry. Section III describes two cross-sectional
data sets that form the basis for the empirical analysis. Section IV
presents estimates of the determinants of wages in the fast food
industry. Section V presents estimates of the differential in fringe
benefits between company-owned and franchised restaurants.
I. FRANCHISING IN THE FAST FOOD INDUSTRY
In a typical franchise agreement, the franchisor (parent
company) grants the franchisee permission to operate a standardThis
content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
78 QUARTERLY JOURNAL OF ECONOMICS
ized restaurant with a specified technology and widely recognized
trademark in exchange for a fixed fee and a monthly royalty on
gross sales (typically 8 percent). In some cases the franchisee is also
required to post an explicit performance bond. The franchisee must
maintain a minimum level of quality and must purchase inputs and
equipment directly from the franchisor or from an approved
supplier. The total start-up cost of a franchised restaurant in a
major chain is normally between $400,000 and $600,000.
The franchise agreement typically requires a commitment by
the franchisee to play an active role in the day-to-day operation,
management, and supervision of the restaurant. Burger King, for
instance, requires an "on-premises" operating owner, while McDonald's,
Wendy's, and Kentucky Fried Chicken do not permit
"absentee" franchisees, and McDonald's will not sell a franchise to
a partnership.3 Franchisees usually choose their restaurant site,
but the franchisor retains the right to veto certain locations. Most
franchise contracts expire after a period of twenty years. When the
franchise contract expires, the parent company may renew the
contract, sell the franchise to another franchisee, or operate the
restaurant itself.
In addition to granting franchise outlets to individual entrepreneurs,
franchisors often own and operate several establishments.
These so-called company-owned units adhere to the same quality
standards established for franchisee-owned units. About 30 percent
of fast food restaurants in the franchise system are companyowned,
but this fraction varies considerably across firms. For
instance, 12 percent of Arby's restaurants, 15 percent of Burger
King restaurants, 18 percent of Kentucky Fried Chicken restaurants,
25 percent of McDonald's restaurants, and 35 percent of
Wendy's restaurants are company-owned.
A. Agency Problems
Agency problems are likely to arise in company-owned outlets
because the parent company cannot perfectly observe the hired
manager's actions and information set (see Ross [1973] and Jensen
and Meckling [1976]). The long-standing practice in the industry
has been to pay managers of company-owned restaurants a fixed
salary that does not directly depend on their performance or their
restaurant's profitability. Although rapid growth in the fast food
3. This information was ascertained from the franchise prospectus of each
aforementioned company.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
OWNERSHIP, AGENCY, AND WAGES 79
industry and accompanying opportunities for promotion may have
provided an incentive for company-owned outlet managers to
provide effort in the past, reports of the recent glut of fast food
restaurants suggest that managers currently have limited opportunity
for upward mobility within fast food chains (see Business
Week [1987] and Emerson [1979]).4
Franchisees have different incentives from company-outlet
managers. Franchisees receive the residual profit generated by
their establishment. The effort franchisees expend managing their
restaurant is reflected in their profit. Consequently, it is reasonable
to expect that franchisees work longer hours and provide more
effort than managers of company-owned outlets.' This aspect of
franchising has been emphasized in much of the relevant industrial
organization literature (for examples, see Caves and Murphy
[1976], Rubin [1978], and Norton [1988]).
The following statement by Fred Turner, President and CEO
of McDonald's, suggests that the corporate leaders of the fast food
industry recognize that the different incentives that managers and
franchisees face lead to a greater outlay of managerial effort at
franchised restaurants. "Running a McDonald's is a three-hundredsixty-three-days-a-year
business and an owner-operator, with his
personal interests and incentives, can inherently do a better job
than a chain manager [Love, 1986, pp. 291-93]."
The discussion so far has assumed that each franchisee
personally manages his or her establishment. Although the restaurant
chains try to select franchisees who will actively and directly
manage their restaurants, in some cases franchisees own more
than one restaurant. Ownership of multiple restaurants will
introduce agency problems in franchised outlets since these franchisees
must hire managers to help run their establishments. Unfortunately,
I do not observe the number of outlets owned by franchisees
in either of my data sets. As a result, possible ownership of multiple
restaurants by individual franchisees will attenuate differences
between company-owned and franchised restaurants.
It should also be noted that franchisors have an incentive to
4. One explanation for the lack of performance-based pay for hired managers is
that until recently the prospect of internal promotion offered a valuable reward for
outstanding performance. The decrease in the likelihood of promotion opportunities
might lead to a change in the structure of compensation of managers in the future.
5. Calvo and Wellisz [1978] present a model of the firm size wage effect that is
very similar to the agency model considered here. Company-owned restaurants are
like large firms in that they have more hierarchical levels.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
80 QUARTERLY JOURNAL OF ECONOMICS
locate company-owned outlets in areas where the cost of monitoring
managers is low to reduce agency problems [Rubin, 1978].
Consistent with this view, evidence suggests that company-owned
restaurants are more likely to be located near headquarters and in
urban areas [Brickley and Dark, 1986]. Such strategic choice over
the location of restaurants also reduces the differential in monitoring
between company-owned and franchised outlets. Furthermore,
since geographic factors have been found to influence labor market
outcomes, it is important to control for city size, population
density, and geographic location in the empirical analysis that
follows.
There is some empirical support for the maintained hypothesis
that employees are less closely monitored and supervised in
company-owned restaurants than in franchisee-owned restaurants.
Table I summarizes the results of a survey of fast food
employees regarding their assessments of supervision at companyowned
and franchised restaurants.6 Forty-five percent of employees
of franchised restaurants strongly agree that the manager of
their restaurant provides adequate supervision to workers, whereas
only 33 percent of company-owned restaurant employees strongly
agree with the same statement. Furthermore, the table shows that
the same pattern of self-reported lower supervision in companyowned
establishments is evident when employees are questioned
about their assistant manager and supervisor. From these tabulations
it is clear that fast food workers report that they receive less
supervision in company-owned outlets than in franchisee-owned
outlets.7
In addition to the subjective opinions of employees, profit data
are consistent with the view that there are substantial agency costs
connected to company-ownership. For instance, Shelton [1967]
examines the profitability of 22 restaurants that experienced a
shift from franchisee-ownership to company-ownership, or vice
versa. In 19 of the 22 cases the restaurant was less profitable under
company ownership than under franchisee ownership. On average,
the profit margin under franchisee ownership was 9.5 percent and
under company management was only 1.8 percent. Because restaurant
location and capital remain unchanged during a change in
6. These survey results are based on the author's tabulations of the National
Institute for Work and Learning's survey of fast food workers. The survey is
described in detail in Section III.
7. These conclusions remain after controlling for individual demographic
characteristics (sex, age, race, etc.) and self-reported satisfaction.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
OWNERSHIP, AGENCY, AND WAGES 81
TABLE I
SUPERVISION IN COMPANY-OWNED AND FRANCHISED RESTAURANTSa
Proportion of employees agreeingb
Company-owned Franchisee-owned
Manager provides adequate supervision 0.326 0.452
to workers (0.010) (0.014)
Assistant manager provides adequate 0.332 0.405
supervision to workers (0.010) (0.013)
Supervisor provides adequate supervi- 0.360 0.468
sion to workers (0.011) (0.014)
Sample size 2,043 1,346
a. Data set is the National Institute of Work and Learning's survey of fast food employees. The data set is
described in Section III. Tabulations were made by the author.
b. Figures represent the proportion of workers who strongly agree with each statement. Standard errors
are shown in parentheses.
ownership, this finding suggests that agency costs substantially
influence the performance of company-owned restaurants.
B. Interior Solution
Franchisors are clearly at an interior solution where some
units are company-owned and others are franchisee-owned. Although
most analyses of franchising emphasize the agency costs
associated with company-ownership, there are several potential
costs to franchising and possible benefits to company-ownership
that lead to the contemporaneous existence of both forms of
ownership within the same franchise chain. Because the forces
that generate this interior solution potentially have implications
for the empirical analysis, they are considered below.
First, franchisees may be forced to bear inefficient risk. If a
franchisee has a substantial proportion of his wealth and income
tied to the performance of a particular restaurant, his investment
portfolio may be insufficiently diversified [Brickley and Dark,
1986]. Furthermore, the franchise contract often places limits on
the franchisee's ability to diversify risks (e.g., no partnerships).
Inefficient risk bearing can lead franchisees to have a higher
required rate of return on their investment, and thus reduce the
franchisor's profit.
Second, a potential disadvantage of franchising is free-riding
by franchisees [Rubin, 1978; Mathewson and Winter, 1985]. For
example, consider a franchised restaurant located in an area where
the probability of serving repeat customers is low. The franchisee
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
82 QUARTERLY JOURNAL OF ECONOMICS
does not bear the full cost of poor quality service and thus has an
incentive to free-ride on the company's reputation by shirking on
quality. On the other hand, company-owned units have less of an
incentive to free-ride because negative externalities reduce the
value of the entire franchise chain. For this reason, Klein and Staft
[1985] argue that franchisors have an incentive to locate companyowned
restaurants in areas where customers have a low probability
of repeat patronage, such as along major highways.8 The free-rider
problem also distorts incentives for other activities, such as
advertising, technological improvement, and product development.
Third, the franchise contract introduces an additional distortion
because it taxes sales, not profits [Ozanne and Hunt, 1971].
The first-best, full-information contract would charge a lump sum
payment or levy a Marshallian profits tax. The practice of charging
a royalty on sales is inefficient because it drives a wedge between
marginal revenue and marginal cost.9
Finally, several observers have argued that an important
explanation for franchising is franchisor capital constraints [Cave
and Murphy, 1976; Ozanne and Hunt, 1971]. According to this
hypothesis, franchisors are capital constrained in their early stages
and sell franchises to raise capital. As the franchisor matures, the
capital constraints are relaxed and company-ownership is possible.
The pattern of growth of most fast food chains supports this view.
For instance, at its inception McDonald's only sold franchise
outlets, but the proportion of domestic company-owned restaurants
has grown rapidly.
C. Wage Determination
With minor deviations most fast food companies follow the
same institutional process of wage determination. Individual franchisees
are autonomous in setting wages and fringe benefits. In
contrast, district and area managers establish the pay rates and
fringe benefits of hourly employees (including assistant managers)
at company-owned outlets. Some portion of the district manager's
pay usually depends on the profitability of the company, but the
restaurant manager is typically paid a fixed annual salary.
The fast food industry is almost entirely non-union. Only one
8. It should be noted, however, that Brickley and Dark [1986] find little
empirical support for this proposition.
9. Cheung [1968] points out that if the royalty is flexible and determined
competitively, a "share" economy will be Pareto efficient.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
OWNERSHIP, AGENCY, AND WAGES 83
fast food restaurant in the United States, a Detroit Burger King, is
currently under a union contract [BNA, 1985]. In addition, few
organizing drives are currently under way or planned. As a result,
wages in the fast food industry are largely unaffected by collective
bargaining or the threat of unionization.
D. Employee Performance
Although fast food employees' jobs are highly routinized and
capital intensive, there remains some scope for employee shirking
and malfeasance. Absenteeism, high turnover, theft, poor service,
waste, and neglect of equipment are examples of detrimental
employee behavior that have a significant effect on profitability.
Greenberger and Steinberg [1986] find that 62 percent of first-time
workers in Orange county gave away goods, falsely claimed to be
sick, stole, damaged property, or worked while intoxicated during
their first nine months of employment. Furthermore, estimates of
the turnover rate of nonmanagement fast food employees run as
high as 300 percent per year, and 81 percent of all establishments
in the industry reported vacancies in the first quarter of 1985
[BNA, 1985].
In comparison with other industries, however, the cost of
shirking in the fast food industry is probably quite low, and the
probability of detecting substandard performance is probably quite
high. A popular book on franchising claims that, "With all the
mechanization, the [fast food] employee has little opportunity to
ruin his product. A bored, half-asleep teenager can perform most
tasks" [Luxenberg, 1985]. To the extent that monitoring is nearly
perfect in the fast food industry regardless of the amount of effort
provided by restaurant managers, the distinction between companyowned
and franchised restaurants will have little effect on employee
performance.
However, one might expect shirking to be more difficult to
detect and to have more costly consequences in relatively skilled
restaurant jobs, such as assistant and shift managers. As a result,
compensation is more likely to be structured to overcome agency
problems for these workers.
II. A SKETCH OF RELEVANT MODELS OF WAGE DETERMINATION
The foregoing discussion suggests that company-owned stores
have greater difficulty monitoring employees relative to stores that
are owned and operated by franchisees. Employee shirking is
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
84 QUARTERLY JOURNAL OF ECONOMICS
therefore more difficult to detect in company-owned stores. Shirking
can take many forms, although the theoretical literature
typically models shirking as unauthorized on-the-job leisure. More
generally, any employee action that reduces output can be thought
of as shirking (e.g., theft or quitting). I begin by describing models
where employees are homogeneous and wages are set to encourage
effort, and then extend the discussion to consider models in which
employees have heterogeneous abilities and characteristics and
wages are set to facilitate selection.
The delayed-payment/bonding model predicts that companyowned
stores will shift relatively more compensation to the end of
the employment contract to discourage shirking. Figure IA illustrates
a hypothetical tenure-earnings profile assuming that company-owned
stores utilize a delayed-payment arrangement to deter
shirking, and that franchisee-owned stores pay employees accordCompany-
Owned
Franchise
o,
0
O T Tenure
FIGURE IA
Hypothetical Tenure-Earnings Profile: Delayed-Payment Model
Company- Owned
a,
War Franchise
0
I 10
Q T Tenure
FIGURE IB
Hypothetical Tenure-Earnings Profile: Efficiency Wage Model
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
OWNERSHIP, AGENCY, AND WAGES 85
ing to their contemporaneous marginal product. When the employee
is first hired, earnings are lower at company-owned stores
than at franchisee-owned stores. However, an employee's wage
rate rises more quickly over time at company-owned stores than at
the franchisee-owned stores. The delayed-payment/bonding model
predicts that the present value of compensation would be equal at
company-owned and franchisee-owned stores. In essence, at the
beginning of employment workers at company-owned firms contribute
toward a bond, and the bond is returned to them if they are not
dismissed for shirking. This contract provides an incentive against
shirking; such an incentive is not necessary at a franchised store
because close supervision by the franchise owner deters shirking.
If company-owned stores are unable to sufficiently backload
compensation to deter shirking (i.e., collect a big enough bond), the
efficiency wage model predicts that they will offer workers a wage
that exceeds their alternative wage. This may be particularly likely
in the fast food industry because the minimum wage may prevent
firms from offering a sufficiently low starting wage.10 In this
situation it will be profit maximizing for company-owned restaurants
to backload compensation to some extent, as well as to offer a
higher present value of compensation over the employee's expected
job duration. Figure IB illustrates such a contract. As in the
delayed-payment/bdnding model, the tenure-earnings profile is
steeper at company-owned stores, but in the efficiency wage model
the present value of compensation is greater at company-owned
stores. This is clear in Figure IB because wages are always higher
at company-owned stores. (Of course, in the efficiency wage model
it might be the case that compensation is initially lower at
company-owned stores.) The key prediction that distinguishes th
two models is that in the efficiency wage model the expected
discounted value of compensation is greater at company-owned
stores than at franchisee-owned stores, whereas in the delayedpayment/bonding
model the discounted value of compensation is
equal at company-owned and franchisee-owned stores.
The discussion thus far has focused on efficiency wage and
delayed-payment/bonding models in which employees are homoge-
10. As Eaton and White [1982] have stressed, workers may also be unwilling to
accept jobs that backload compensation for fear that firms will dismiss them prior to
the date when compensation exceeds their marginal product, or employees may be
liquidity constrained so they prefer to receive compensation earlier rather than
later.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
86 QUARTERLY JOURNAL OF ECONOMICS
neous and have some discretion over their work effort. In another
class of models that also may be relevant in the fast food industry,
firms may structure compensation to facilitate the selection of
high-productivity employees. In particular, it is probably the case
that managers of company-owned stores have less of an incentive
to carefully recruit and screen job applicants compared with
franchise owners. If workers are heterogeneous in terms of their
productive abilities and propensity to quit, franchisees will more
effectively screen applicants to identify those who are most productive.
Models by Guasch and Weiss [1981], Weiss [1980], and others,
predict that in this situation company-owned stores will offer a
relatively high wage to improve the pool of job applicants. Thus,
another reason for company-owned stores to pay higher wages
than franchisee-owned stores do might be to improve the pool of
applicants.
Finally, the relationship between turnover and pay deserves
special emphasis in an industry like the fast food industry, which
has an extremely high rate of turnover. Turnover is costly to
restaurants because resources are spent recruiting, screening, and
training workers. For reasons discussed above, turnover is likely to
be more costly for company-owned stores. Therefore, companyowned
stores have a greater incentive to design compensation in a
way to attract workers who are less prone to quit, and to discourage
current workers from quitting. If workers differ in their propensity
to quit, Salop and Salop [1976] have argued that establishments
that have a high cost of turnover will provide relatively steep
tenure earnings profiles to discourage quit-prone employees from
applying, as shown in Figure IA. Furthermore, even if all applicants
have an equal quit propensity, company-owned stores could
reduce turnover by paying more generous compensation than
franchisee-owned stores do (see Salop [1979]).
On the basis of the estimated earnings profiles, it is not
possible to distinguish among the competing explanations for why
firms tilt the tenure-earnings profile or offer relatively high wages.
Nonetheless, evidence of higher discounted compensation at company-owned
stores would support the class of efficiency wage
models, whereas evidence of steeper tenure-earnings profiles at
company-owned stores but equal discounted compensation would
support the delayed-payment/bonding model. And evidence of an
identical wage structure in company-owned and franchised establishments
would cast doubt on the importance of both types of
incentive models.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
OWNERSHIP, AGENCY, AND WAGES 87
III. DATA
Two cross-sectional data sets on the fast food industry are
used in the empirical work. The primary data set was collected by
the National Institute for Work and Learning (NIWL).'1 During
the fall of 1982 the NIWL mailed questionnaires to 7,000 workers
employed by a random sample of 273 fast food restaurants. The
questionnaire contained a rich set of questions about the employees'
job requirements, human capital, wages, and fringe benefits.
Employee responses were merged to company-provided data on the
form of ownership, location, and size of the restaurant. Each
employee was paid $5 upon completing the questionnaire, and two
follow-up questionnaires were sent to nonrespondents. The overall
survey response rate was 66 percent.
The samples analyzed in this paper contain either crew
workers (part- or full-time) or assistant and shift managers who
are employed by companies that utilize both company and franchisee
ownership arrangements. All workers are hourly, nonexempt
employees. Four companies in the survey-McDonald's, Kentucky
Fried Chicken, Arby's, and Roy Rogers-use both types of ownership
arrangements. The final samples contain 1,889 nonmanagement
workers and 198 hourly assistant and shift managers from
204 restaurants.
Means and standard deviations for the sample are presented in
Table II. In general, the workers at company-owned and franchised
restaurants appear remarkably similar. For instance, employees of
company-owned and franchised restaurants have almost an identical
average number of years of schooling and high school grade
point average. At least on the basis of these observable characteristics,
there is little evidence of differential sorting of employees
between company-owned and franchisee-owned outlets.
The means also indicate that compared with franchised outlets,
company-owned outlets are located in larger cities and in cities
that have slightly higher average wages. In addition, companyowned
restaurants tend to employ fewer workers than franchised
restaurants.' The latter finding is consistent with Norton's [1988]
argument that franchising is the preferred form of ownership for
larger-sized establishments because large organizations are inher-
11. See Charner and Fraser [19841 for a thorough description of the survey
design and a copy of the questionnaire.
12. In the second data set I also find that on average franchised stores are
larger than company-owned stores.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
88 QUARTERLY JOURNAL OF ECONOMICS
TABLE II
MEANS AND STANDARD DEVIATIONS OF THE NIWL DATA SET,
BY TYPE OF OWNERSHIP
Sample
Franchisee- Company
Variable All owned owned
Asst. and shift mgr. hourly 4.59 4.35 4.75
wage (1.03) (0.71) (1.17)
Crew worker hourly wage 3.59 3.57 3.61
(0.35) (0.31) (0.38)
Years of education 11.99 11.95 12.01
(1.40) (1.39) (1.40)
Tenure 1.56 1.51 1.59
(1.35) (1.25) (1.41)
Age 20.10 20.13 20.08
(5.57) (5.80) (5.41)
GPA 84.59 84.81 84.43
(5.90) (5.84) (5.94)
Proportion part-time 0.64 0.67 0.62
(0.48) (0.47) (0.48)
College prep. high school 0.46 0.46 0.46
(0.50) (0.50) (0.50)
Proportion black 0.13 0.08 0.16
(0.33) (0.27) (0.36)
Proportion hispanic or other 0.07 0.08 0.07
race (0.26) (0.27) (0.25)
Proportion female 0.64 0.66 0.63
(0.48) (0.47) (0'48)
Proportion student 0.52 0.53 0.51
(0.50) (0.50) (0.50)
Round-trip commute time 14.22 14.46 14.06
(9.45) (9.37) (9.50)
Establishment size 39.80 46.11 35.43
(27.10) (38.07) (13.97)
Log local annual wage 9.62 9.60 9.65
(0.15) (0.15) (0.14)
Log city size 11.69 11.53 11.80
(1.78) (1.78) (1.76)
Log population density 8.26 8.26 8.27
(0.76) (0.81) (0.73)
Sample size 2,087 852 1,235
ently more difficult to monitor and therefore benefit more from
diligent management.
On average, low-level managers earn 40 cents more per hour at
company-owned outlets than at franchised outlets, while crew
workers earn just 4 cents more per hour at company-owned outlets
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
OWNERSHIP, AGENCY, AND WAGES 89
than at franchised outlets. For low-level restaurant managers, the
average hourly wage rate is $4.59 with a coefficient of variation of
22 percent, while the mean hourly wage rate for crew workers is
$3.59 with a coefficient of variation of less than 10 percent. Less
than 20 percent of employees in the sample are paid the minimum
wage, and 2 percent are paid less than the legal minimum.
Combining all employees, the average job tenure is about one
month greater at company-owned outlets. The overall tenure
differential is not statistically significant at the 5 percent level.
However, low-level managers employed by company-owned outlets
have a statistically significant additional half year of job tenure
than workers in equivalent jobs at franchised outlets. One would
expect to observe longer job tenures at company-owned outlets if
the wage differentials noted above represented true economic
rents.
Bureau of National Affairs Data Set
The second data set was collected by the Bureau of National
Affairs (BNA) in 1985. The BNA surveyed the managers of 108 fast
food restaurants in five metropolitan areas and received responses
from 47 restaurants. Restaurants in the Burger King, Kentucky
Fried Chicken, McDonald's, and Wendy's chains were surveyed.
The survey contains questions on the starting wage rate of full- and
part-time nonmanagement employees, restaurant location, and
restaurant size. Unlike the NIWL survey, information is not
available for assistant and shift managers. Despite this limitation
and its small sample size, the BNA survey has an important
advantage over the NIWL survey: it allows one to control for the
city in which the restaurant is located because there are observations
on several restaurants in the same cities. In the estimates
below, this is accomplished by including a set of city dummy
variables.
IV. EMPIRICAL RESULTS
For each occupational group the NIWL data set is used to
estimate parameters of a human capital earnings function of the
form:
(1) In Wj = Xlj ? C,5 + p4, i = (1,.. .,n); j = (1, ... mi
where Wi, is the hourly wage rate for workerj at restaurant i, Xi, is a
vector of independent variables including tenure and three dummy
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
90 QUARTERLY JOURNAL OF ECONOMICS
variables indicating the parent company, CQ is a d
indicating whether a restaurant is company-owne
term, and 13 and 8 are parameters to be estimated. In some
specifications the company-ownership dummy variable is also
interacted with firm-specific tenure to test for the presence of
delayed-payment contracts.'3 There are n firms in the sample and
mi workers at firm i.
Because in many cases the NIWL data set contains observations
on multiple employees from the same establishment (m, > 1),
an error components structure is assumed. Specifically, I assume
that LIp consists of a firm-specific component (aX) and a workerspecific
component (Ej) with the following properties:
(2) Rj Ci + Eij.
(3) E (oi) E(Eij) =c E (?,Ei,) = 0
(4) a2 E(c4'); o = E(E2J) for all i and j.
A random effects model is estimated because ordinary least squares
is inefficient and yields inconsistent standard errors under these
assumptions.14
Table III reports earnings regressions for three separate
samples of fast food workers: assistant and shift managers, fulltime
crew workers, and part-time crew workers. Although an
objection can be raised that assistant managers perform different
functions at company-owned and franchised restaurants, it should
be noted that the regressions control for 11 detailed job tasks.
Furthermore, in both types of restaurants these low-level managers
are hourly, nonexempt employees.
First, consider the results for assistant and shift managers. In
column 1 the wage differential associated with company-ownership
is constrained to an intercept shift. The estimated wage differential
associated with working at a company-owned restaurant is 8.7
13. A Chow test of the hypothesis that it is appropriate to pool observations
from company-owned and franchised restaurants allowing for an intercept change
and differential return to tenure is not rejected at the 5 percent level for each
occupation. Similarly, one would not reject a test of the appropriateness of pooling
workers from different parent companies after allowing for separate company
intercepts. Consequently, a single equation with company dummies is estimated.
14. See Judge et al. [1985] and Johnston [19841 for a discussion of random
effects estimation. The estimator used here differs slightly from the textbook case
because the panel is unbalanced. The estimates and conclusions are not qualitatively
different when the equations are estimated by OLS. A table containing the
OLS results is available on request.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
TABLE III
RANDOM EFFECTS ESTIMATES OF THE DETERMINANTS OF WAGES IN THE
FAST FOOD INDUSTRY
(DEPENDENT VARIABLE: LOG HOURLY WAGEa)
Sample
Independent Assistant and Full-time Part-time
Variables shift managers crew workers crew workers
(1) (2) (3) (4) (5) (6)
Intercept -0.411 -0.114 -0.115 -0.146 0.637 0.651
(1.007) (0.975) (0.298) (0.311) (0.191) (0.193)
Company-owned (1 = yes) 0.083 0.013 0.017 -0.015 0.005 -0.002
(0.038) (0.054) (0.009) (0.013) (0.005) (0.007)
Company-owned x tenure 0.025 - 0.020 0.006
(0.014) (0.006) (0.003)
Tenure 0.042 0.025 0.032 0.020 0.028 0.024
(0.007) (0.012) (0.003) (0.005) (0.002) (0.002)
Education 0.011 0.010 0.005 0.005 0.005 0.005
(0.008) (0.008) (0.003) (0.003) (0.001) (0.001)
GPA . 100 0.027 0.025 0.057 0.073 0.068 0.066
(0.220) (0.221) (0.070) (0.069) (0.029) (0.029)
College prep. high school 0.017 0.012 0.003 0.006 0.003 0.003
curriculum (0.024) (0.025) (0.009) (0.009) (0.003) (0.003)
Student -0.053 -0.055 -0.022 -0.022 -0.006 -0.006
(0.029) (0.029) (0.009) (0.009) (0.004) (0.004)
Age 0.004 0.006 0.001 0.002 0.0011 0.0011
(0.002) (0.002) (0.001) (0.001) (0.0004) (0.0004)
Married 0.025 0.013 0.034 0.033 0.006 0.007
(0.032) (0.033) (0.012) (0.012) (0.006) (0.006)
Black -0.087 -0.081 -0.015 -0.015 -0.010 -0.011
(0.041) (0.041) (0.011) (0.011) (0.006) (0.006)
Hispanic and other -0.000 0.006 -0.009 -0.008 -0.003 -0.003
(0.046) (0.046) (0.015) (0.015) (0.006) (0.006)
Female -0.029 -0.029 -0.006 -0.006 -0.010 -0.010
(0.025) (0.025) (0.011) (0.011) (0.005) (0.005)
Round-trip commute time . 0.378 0.353 0.071 0.076 -0.021 -0.021
100 (0.104) (0.105) (0.040) (0.040) (0.017) (0.017)
Weekly hours 0.003 0.003 0.0005 0.0007 0.0005 0.0005
(0.001) (0.001) (0.0007) (0.0007) (0.0002) (0.0002)
Log establishment size 0.009 0.019 0.011 0.003 0.003 0.003
(0.038) (0.038) (0.012) (0.012) (0.007) (0.007)
Loglocal annual wage 0.162 0.134 0.127 0.127 0.044 0.043
(0.113) (0.109) (0.033) (0.035) (0.021) (0.022)
Log city size 0.016 0.019 -0.008 -0.009 -0.003 -0.003
(0.013) (0.012) (0.003) (0.004) (0.007) (0.002)
Log population density -0.057 -0.062 0.012 0.012 0.003 0.003
(0.030) (0.030) (0.009) (0.009) (0.005) (0.005)
Assistant managerb 0.178 0.177
(0.064) (0.062)
Job task dummies (11)Y Yes Yes Yes Yes Yes Yes
Parent company dummies (3) Yes Yes Yes Yes Yes Yes
Census region dummies (8) Yes Yes Yes Yes Yes Yes
U2 0.012 0.012 0.007 0.007 0.003 0.003
o2 0.005 0.004 0.001 0.001 0.001 0.001
a. Data set is NIWL survey of fast food employees. S
198 for columns 1 and 2,615 for columns 3 and 4, and 1,274 for columns 5 and 6.
b. Dummy variable that equals one if the worker is an assistant manager and zero if he is a shift manager.
All assistant and shift managers are hourly, nonexempt employees.
c. Job tasks that are performed all or most of the time are coded 1; job tasks that are performed only
sometimes, seldom, or never are coded 0. The eleven job tasks include cooking, preparing food (noncooking),
packing orders, taking orders, handling money, hosting dining area, suggestive selling, unloading trucks,
cleaning the restaurant, cleaning equipment, and training workers.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
92 QUARTERLY JOURNAL OF ECONOMICS
percent, with a t-ratio of 2.31.15 To put the magnitude of this wage
differential in perspective, note that the standard deviation in log
wages among these workers is 0.19, so the company-ownership
wage differential equals nearly half the interworker dispersion in
earnings in the occupation.
Column 2 presents an estimate of the same specification
adding an interaction between tenure and company-ownership.
These results indicate that low-level managers' wages grow more
than twice as quickly over time at company-owned outlets than at
franchised outlets.16 Furthermore, the company-ownership dummy
variable becomes statistically insignificant and close to zero when
the interaction term is included, which implies that starting pay is
roughly the same at company-owned and franchised outlets. In
sum, the wage profile for assistant managers approximates the
hypothetical profile depicted in Figure 1B.17
Next consider the estimates for nonmanagement workers.
Columns 3 and 5 show a small positive wage differential associated
with working at company-owned establishments. Full-time workers
earn 1.7 percent greater wages at company-owned restaurants
than at franchisee-owned restaurants, while for part-time workers
the company-ownership wage differential is just 0.5 percent.
Although these coefficients are precisely estimated, they are trivial
by most economic standards.
In columns 4 and 6 an interaction between job tenure and the
company-ownership dummy variable is added to the independent
variables. As was the case with low-level managers, the interaction
term has a positive and statistically significant effect on wages for
nonmanagement workers. This also supports the conclusion that
earnings grow more rapidly at company-owned outlets than at
franchisee-owned outlets. In addition, this specification finds that
15. The percentage differential is calculated from Table III as (exp (0.083) -
1) x 100.
16. Results are qualitatively unchanged when the square of tenure and its
interaction with the company-owned dummy variable are added to the regression.
17. Interestingly, on average the returns to firm-specific tenure are about the
same magnitude in the fast food industry as researchers have found for the economy
as a whole, but the tenure-earnings profile is flatter for fast food workers than for a
random sample of young workers. For example, Mincer and Jovanovic [1981]
estimate that the return to the first year of job tenure is 6.5 percent for young men
and 3 percent for all men. I also note that the estimated tenure-earnings profile for
fast food workers is less prone to criticism along the lines raised by Altonji and
Shakotko [1987] because jobs and firms in the industry are homogeneous and
because workers have little opportunity to sort among firms given their short
experience in the labor market.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
OWNERSHIP, AGENCY, AND WAGES 93
the company-owned shift dummy variable is negative but statistically
insignificant for each group of workers. Again, these tenureearnings
profiles are similar to the hypothetical efficiency wage
model depicted in Figure IB, but the difference between the two
profiles for nonmanagement workers is quite small and of little
practical significance.
Finally, consider the other variables in the wage regressions.
The human capital variables-education, age, grade point average,
type of high school curriculum, 11 job task dummy variables, and a
dummy variable indicating whether the worker is a studentmostly
have their expected signs and are usually statistically
significant. These observed human capital variables, however, do
not have a very sizable effect on wages in the fast food industry.
Similarly, the wage differential for married workers is positive but
smaller than is typically observed in other industries (see Korenman
and Neumark [1987]).
The wage differentials observed for race and gender in the
sample of nonmanagement fast food workers are also small when
compared with those in other industries. Minority and female
nonmanagement workers earn from 1 to 2 percent lower wages
than white and male employees, all else constant. In the sample of
low-level managers, however, the differentials are larger. Black
assistant and shift managers earn about 8 percent lower wages
than whites in similar positions, while females earn about 3
percent lower wages than males. The finding of a relatively modest
wage gap for women is particularly interesting given the historical
pattern of female employment in the fast food industry. McDonald's
Corporation, for instance, officially refused to hire women
until 1968, but women currently make up 57 percent of McDonald's
crew workers (see Love [1986, p. 294]).
The variable round-trip commute time (measured in minutes)
is included in the regressions as a working condition control. The
theory of equalizing differences predicts that a job located in an
area that requires a long commute to work will offer a compensating
wage premium to attract workers. Furthermore, since there are
systematic geographic differences between company-owned and
franchised restaurants that affect travel time to work (e.g., company-owned
restaurants located near highways), omission of this
variable could bias the company-owned wage differential. A 30minute
increase in commuting time is associated with an 11
percent increase in the wage rate for low-level managers and a 2
percent increase for full-time crew workers.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
94 QUARTERLY JOURNAL OF ECONOMICS
It has been noted that franchisors have an incentive to locate
company-owned units in urban areas to economize on transaction
costs [Rubin, 1978]. Furthermore, Fuchs [1967] and Hoch [1972]
document that wages increase with city size. The log of the
population of the city where the restaurant is located and the log of
the population density are included as regressors to control for
these locational factors.'8 To further control for differences in wage
levels across areas, the equations include the log of the average
annual salary in the county in which the restaurant is located and
eight region dummy variables. In particular, the estimates indicate
that fast food restaurant workers' wages are sensitive to the local
wage rate.
A large empirical literature finds that wages rise with establishment
size (for examples see Lester [1967], Masters [1969], Mellow
[1982], and Brown and Medoff [1989]). In the fast food industry,
however, the estimated establishment size-wage effect is close to
zero and statistically insignificant in each specification and sample.
Moreover, this result is qualitatively unchanged when five restaurant
size-class dummy variables are used as the establishment size
measure to allow for flexibility in the establishment size-wage
effect. The finding of an insignificant effect of restaurant size on
wages in the fast food industry takes on added significance in view
of the finding that on average franchised restaurants are larger
than company-owned restaurants.
A. Analysis of the BNA Data Set
The wage question in the 1985 BNA survey pertains to the
starting wage rate. This is a severe limitation because many
restaurants were constrained to pay the legal minimum wage to
newly hired workers. For nearly 75 percent of the restaurants in
the sample the starting wage rate of part-time workers equals the
minimum wage, while for 47 percent of the restaurants the
starting wage rate of full-time workers equals the minimum wage.
18. The average local wage, city size, and population density variables were
merged to the NIWL data set by the author. Cities are identified from the
restaurant's zip code and the city size and population density variables are derived
from 1980 Census data reported in The County and City Data Book, 1983
(Washington, DC: Government Printing Office, 1983). The average annual local
wage is the average wage in the county, and is drawn from U. S. Bureau of the
Census, County Business Patterns 1981, Table 1E.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
OWNERSHIP, AGENCY, AND WAGES 95
Since the meaningful variation in these wage data is whether the
restaurant initially pays the minimum wage or more than the
minimum wage, I shall focus on this decision.
An examination of the unconditional means shows that franchisee-owned
restaurants are more likely than company-owned
restaurants to pay the minimum wage to newly hired employees.
More than 90 percent of franchisee-owned restaurants start their
new part-time employees at the minimum wage, while 60 percent
of company-owned restaurants pay the legal minimum to new
employees. For full-time workers the comparable figures are 55
percent at franchisee-operated outlets and 40 percent at companyowned
outlets.
Table IV presents probit equations of the likelihood that a
restaurant pays the minimum wage to new hires. The estimates
show a substantial and statistically significant difference in the
probability that company-owned outlets pay the minimum wage to
newly hired part-time workers, holding constant restaurant size
and location. The company-owned differential for full-time workers,
however, is much smaller and statistically insignificant. The
coefficient on the company-ownership dummy implies that compared
with franchised outlets, company-owned outlets are 45
TABLE IV
PROBIT ESTIMATES OF PROBABILITY OF PAYING THE MINIMUM WAGEa
Sample
coefficient (SE)
Independent Mean
variables (SD) Full-time Part-time
Intercept 1.00 3.207 4.642
(0.00) (2.831) (3.296)
Company-owned restaurant 0.53 0.183 -1.396
(0.50) (0.483) (0.586)
Log estab. size 3.71 0.881 -0.694
(0.64) (0.730) (0.828)
City dummies (4) Yes Yes
Parent company dummies (3) Yes Yes
Log likelihood fn. -26.44 -20.36
a. The dependent variable equals one if the starting wage equals the minimum wage, and zero if the starting
wage exceeds the minimum wage. Sample size is 47. Data are from BNA survey of establishments. The
proportion of restaurants that pay the minimum wage to newly hired part-time workers is 0.745, and the
proportion of restaurants that pay the minimum wage to newly hired full-time workers is 0.468.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
96 QUARTERLY JOURNAL OF ECONOMICS
percentage points less likely to start workers at the minimum
wage. 19
These results are particularly relevant in light of the delayedpayment/bonding
model, which predicts that company-owned restaurants
would initially pay lower wages than franchised restaurants
as part of an implicit performance bond. Since it could be
argued that in the fast food industry the minimum wage is often a
binding constraint on starting pay, one might expect that companyowned
outlets are constrained in their ability to tilt the tenureearnings
profile. An implication of this view is that company-owned
outlets would lower initial compensation and backload compensation
to the extent possible. In contrast, the results indicate that the
minimum wage is a greater constraint on franchisee-run outlets
than on company-run outlets, which suggests that company-owned
outlets do not lower starting pay to the extent possible. This
finding, together with the results in Table III, casts doubt on the
bonding model.
V. FRINGE BENEFITS
To the extent that fringe benefits augment or offset the
observed company-owned wage effect, differences in total compensation
will be affected. Table V examines the effect of companyownership
on several fringe benefits for nonmanagement workers.20
Unfortunately, the dollar value of fringe benefits is not
available. The dependent variable equals one if the employee
receives a fringe benefit and zero if he or she does not. Probit
equations are estimated to control for human capital, job tasks, and
demographic factors.
The results indicate that the provision of fringe benefits is
more generous in company-owned establishments than in franchisee-owned
establishments. For instance, employees at companyowned
restaurants are 43 percentage points more likely to receive
free meals, 24 percentage points more likely to have a paid
vacation, and 7 percentage points more likely to have paid holidays
19. This difference was calculated as 13 +(z), where 13 is the coefficient on the
company-owned dummy from the probit equation, +( ) is the standard normal
density, z = V-`(p), dP-'( ) is the inverse cumulative normal distribution
function, andp is the proportion of restaurants in the sample that start workers at
the minimum wage.
20. Since there were not important differences between part- and full-time
workers as far as fringe benefits are concerned, they are combined into one sample.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
OWNERSHIP, AGENCY, AND WAGES 97
TABLE V
THE EFFECT OF COMPANY-OWNERSHIP ON FRINGE BENEFITS FOR
NONMANAGEMENT WORKERS
PROBIT ESTIMATESa
(3)
Proportionate
(2) differential
(1) Company-owned between
Fringe benefit dummy variable company-owned and
(mean) (standard error) franchised unitsb
1. Free meals 1.312 0.427
(0.673) (0.086)
2. Uniform allowance 0.172 0.031
(0.114) (0.093)
3. Paid vacation 0.611 0.239
(0.467) (0.075)
4. Paid sick leave 0.488 0.024
(0.032) (0.163)
5. Paid holidays 0.532 0.071
(0.105) (0.112)
6.Insurance benefits 0.359 0.094
(0.168) (0.089)
a. Other independent variables are age, race dummies (2), sex, education, GPA, college prep. dummy,
student dummy, weekly hours, job task dummies (11), job tenure and its square, commute time, census region
dummies (9), log restaurant size, log city size, parent company dummies (3), and an intercept. Differentials
represent the difference in the probability of receiving a fringe benefit between company-owned and franchised
establishments for workers with the average value of all independent variables.
b. Sample size is 1,880.
than comparable employees in franchisee-owned restaurants. In
addition, uniform allowances, paid sick leave, and insurance benefits
are more likely to be provided in company-owned restaurants
than in franchisee-owned restaurants.
On average, the provision of fringe benefits is 14.8 percent
points more likely by company-owned establishments than by
franchisee-owned establishments.2 Since voluntary fringe benefits
typically make up 15 percent of total compensation in eating and
drinking establishments (National Income and Product Accounts),
the fringe benefit differential increases the differential in total
compensation between company-owned and franchised restau-
21. A weighted average results in a 16 percent fringe benefit differential,
assuming that free meals, paid vacations, paid holidays, and insurance benefits each
account for 20 percent of fringe benefits, and that uniform allowance and paid sick
leave each account for 10 percent of fringe benefits.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
98 QUARTERLY JOURNAL OF ECONOMICS
rants by about two percentage points. Finally, it should be emphasized
that-unlike the finding for wages-there is no significant
evidence of a comparatively steeper tenure-fringe benefit profile at
company-owned restaurants when an interaction term for tenure
and company-ownership is added to the equations.
VI. SUMMARY AND CONCLUSION
A core prediction of the efficiency wage model is that wages and
monitoring costs are positively correlated. On the other hand, the
delayed-payment/bonding model predicts that a firm will tilt its
workers' tenure-earnings profile without changing the present
value of wages in response to costly monitoring. The institutional
arrangements of the franchise system provide a natural experiment
to test these predictions because principal-agent relationships
are likely to cause incentives for inferior monitoring and
supervision in company-owned outlets relative to franchiseeowned
outlets. Consequently, this paper uses two cross-sectional
surveys of the fast food industry to test the effect of organizational
structure on pay.
The main empirical results may be summarized as follows:
First, employees of company-owned and franchised restaurants
have very similar characteristics. Second, wages are about 9
percent higher for low-level managers, and 1 to 2 percent higher for
crew workers who are employed by company-owned outlets rather
than by franchisee-owned outlets. Third, the tenure-earnings
profile is steeper at company-owned restaurants than at franchiseeowned
restaurants. Assuming a 5 percent interest rate and the
average level of job tenure, the point estimates imply that the
present value of earnings is about $1,250 greater for assistant and
shift managers, and about $75 greater for full-time workers at
company-owned outlets relative to franchisee-owned outlets (in
1982 dollars). Moreover, accounting for fringe benefits increases
the differential in compensation.
The empirical evidence on the wage structure of low-level
managers in the fast food industry provides some support for the
view that wages are adjusted to overcome incentive problems. Since
workers earn greater compensation where monitoring is more lax,
the results are consistent with the efficiency wage model. Moreover,
the finding that wages grow more rapidly over time at
company-owned units suggests that to some extent compensation
is backloaded to provide an incentive against shirking. For crew
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
OWNERSHIP, AGENCY, AND WAGES 99
workers, however, differences in wages between company-owned
and franchised units are trivial, suggesting that their compensation
is not noticeably affected by monitoring problems. Although
the results for low-level managers provide some support for the
incentive contracts literature, the estimates for crew workers
suggest that incentive contracts are not very important empirically
for this group.
An alternative interpretation of the modest wage differentials
found here is that differences in the level and growth of fast food
workers' earnings reflect rent sharing at company-owned outlets.
It could be argued that managers of company-owned outlets do not
personally bear the full cost of paying higher wages since they do
not have a residual claim on the restaurant's profit, while franchisees
have strong incentives to minimize costs. Furthermore, one
might expect that managers of company-owned outlets have a
motive to pay supra-competitive wages because they like their
workers, or because they find their jobs less onerous if their
workers are well paid.
There are two factors, however, that weigh against this
expense-preference/rent sharing interpretation of the empirical
results. First, fast food workers' wages are typically set at the
district level by middle-level managers instead of by restaurant
managers who have daily contact with workers. Consequently, the
actual wage setters probably do not benefit directly by sharing
rents with restaurant workers. And second, the expense-preference
model does not easily explain the steeper tenure-earnings profile at
company-owned outlets. Nonetheless, it is conceivable that agency
problems lead company-owned outlets to err on the side of
generosity in setting wages, while enterprising franchises diligently
minimize costs.
An important question is whether the findings in this paper
can be generalized to other employment settings. Several factors
suggest that efficiency wages are more likely to surface as a worker
discipline device in the fast food industry than in other industries.
Fast food jobs have short durations with little chance of advancement
through internal job ladders. Furthermore, the minimum
wage may constrain delayed-payment/bonding contracts in this
industry, especially for crew workers. On the other hand, the high
degree of monitoring in fast food restaurants (e.g., by customers
and equipment) regardless of the form of ownership reduces the
need to induce effort through alternative means. As a result, a
worthwhile extension of the present paper would be a study of the
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
100 QUARTERLY JOURNAL OF ECONOMICS
effect of franchise ownership on wages in industries where employee
shirking is more difficult to monitor, such as the gasoline
service industry or the hotel industry.
PRINCETON UNIVERSITY AND
NATIONAL BUREAU OF ECONOMIC RESEARCH
REFERENCES
Akerlof, George, and Lawrence Katz, "Workers' Trust Funds and the Logic of Wage
Profiles," Quarterly Journal of Economics, CIV (1989), 525-36.
Altonji, Joseph, and Robert Shakotko, "Do Wages Rise with Job Seniority?" Review
of Economic Studies, LIV (1987), 437-59.
Becker, Gary, and George Stigler, "Law Enforcement, Malfeasance, and the
Compensation of Enforcers," Journal of Legal Studies, III (1974), 1-18.
Brickley James, and Frederick Dark, "The Choice of Organizational Form: The
Case of Franchising," Journal of Financial Economics, XVIII (1986), 197-218.
Brown, Charles, and James Medoff, "The Employer Size-Wage Effect," Journal of
Political Economy, XCVII (1989), 1027-58.
Bulow, Jeremy, and Lawrence Summers, "A Theory of Dual Labor Markets with
Application to Industrial Policy, Discrimination, and Keynesian
Unemployment," Journal of Labor Economics, IV (1986), 376-414.
Bureau of National Affairs, "Employee Relations in the Fast Food Industry," in
Retail/Services Labor Report, BNA, June 10, 1985.
Business Week, "McWorld?" (October 13, 1987), 78-86.
Calvo, Guillermo, and Stanislaw Wellisz, "Supervision, Loss of Control, and the
Optimum Size of the Firm," Journal of Political Economy, LXXXVI (1978),
943-52.
Cappelli, Peter, and Keith Chauvin, "An Inter-Plant Test of Efficiency Wage
Premiums," mimeo, University of Pennsylvania, 1988.
Caves, Richard, and William Murphy, II, "Franchising: Firms, Markets, and
Intangible Assets," Southern Economic Journal, XLII (1976), 572-86.
Charner, Ivan, and Bryna S. Fraser, Fast Food Jobs (Washington, DC: The National
Institute for Work and Learning, 1984).
Cheung, Steven, "Private Property Rights and Sharecropping," Journal of Political
Economy, LXXVI (1968), 1107-22.
Dickens, Williams, Lawrence Katz, Kevin Lang, and Lawrence Summers, "Employee
Crime and the Monitoring Puzzle," Journal of Labor Economics, VII
(1989),331-48.
Eaton, C., and W. D. White, "Agent Compensation and the Limits of Bonding,"
Economic Inquiry, XX (1982), 330-43.
Emerson, Robert, Fast Food: The Endless Shakeout (New York: Lebhar Friedman
Books, 1979).
Fuchs, Victor, "Differentials in Hourly Earnings by Region and City Size, 1959,"
NBER, Occasional Working Paper No. 101, 1967.
Greenberger, Ellen, and Laurence Steinberg, When Teenagers Work: The Psychological
and Sociological Costs of Adolescent Employment (New York: Basic Books,
1986).
Groshen, Erica, and Alan Krueger, The Structure of Pay and Supervision in
Hospitals," Industrial and Labor Relations Review, XLIII (1990), 134-46.
Guasch, J. L., and Andrew Weiss, "Self-Selection in the Labor Market," American
Economic Review, LXXI (1981), 275-84.
Hoch, Irving, "Income and City Size," Urban Studies, (1972), 299-327.
Hutchens, Robert, "Delayed Payment Contracts and a Firm's Propensity to Hire
Older Workers," Journal of Labor Economics, IV (1986), 439-57.
Jensen, Michael, and William Meckling, "Theory of the Firm: Managerial Behavior,
Agency Costs and Ownership Structure," Journal of Financial Economics, III
(1976), 305-60.
Johnston, J., Econometric Methods (New York: McGraw-Hill Book Company,
1984), pp. 396-407.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms
OWNERSHIP, AGENCY, AND WAGES 101
Judge, George, W. E. Griffiths, R. Hill, Helmut Lutkerpohl,
and Practice of Econometrics 2nd ed. (New York: John Wiley and Sons, 1985),
pp. 515-45.
Klein, Benjamin, and L. F. Staft, "The Law and Economics of Franchise Tying
Contracts," Journal of Law and Economics, XXIIX (1985), 345-61.
Korenman, Sanders, and David Neumark, "Does Marital Status Really Affect
Wages," unpublished, Harvard University, September 1987.
Lazear, Edward, "Agency, Earnings Profiles, Productivity and Hours Restrictions,"
American Economic Review, LXXI (1981), 606-20.
, and Robert Moore, "Incentives, Productivity, and Labor Contracts," Quarterly
Journal of Economics, IC (1984), 275-96.
Leonard, Jonathan, "Carrots and Sticks: Pay Supervision, and Turnover," Journal
of Labor Economics, V (1987), S136-S153.
Lester, Richard, "Pay Differentials by Size of Establishment," Industrial Relations,
VII (1967), 57-67.
Love, John, McDonald's: Behind the Arches (New York: Bantam Books, 1986).
Luxenberg, Stan, Roadside Empires (New York: Penguin Books, 1985).
Mathewson, G. F., and Ralph Winter, "The Economics of Franchise Contracts,"
Journal of Law and Economics, XXVIII (1985), 503-26.
Masters, Stanley, "An Interindustry Analysis of Wages and Plant Size," Review of
Economics and Statistics, LI (1969), 341-45.
Mellow, Wesley, "Employer Size and Wages," Review of Economics and Statistics,
LXIV (1982), 495-501.
Mincer, Jacob, and Boyan Jovanovic, "Labor Mobility and Wages," in Studies in
Labor Markets, Sherwin Rosen, ed. (Chicago: University of Chicago Press,
1981), pp. 21-64.
Norton, Seth, "An Empirical Look at Franchising as an Organizational Form,"
Journal of Business, LXI (1988), 197-218.
Ozanne, Urban, and Shelby Hunt, The Economic Effects of Franchising (Washington,
DC: GPO, Select Committee on Small Business, United States Senate, 92
Congress, 1971).
Ross, Stephen, "The Economic Theory of Agency: The Principal's Problem,"
American Economic Review, Proceedings, LXIII (1973), 134-39.
Rubin, P. H., "The Theory of the Firm and the Structure of the Franchise
Contract," Journal of Law and Economics, XXI (1978), 223-33.
Salop, Joanne, and Steven Salop, "Self-Selection and Turnover in the Labor
Market," Quarterly Journal of Economics, LXXXXI (1976), 619-27.
Salop, Steven, "A Model for the Natural Rate of Unemployment," American
Economic Review, LXIX (1979), 117-25.
Shapiro, Carl, and Joseph Stiglitz, "Involuntary Unemployment as a Worker
Discipline Device," American Economic Review, LXXIV (1984), 433-44.
Shelton, J. P., "Allocative Efficiency vs. 'X-Efficiency': Comment," American
Economic Review, LVII (1967), 1252-58.
Weiss, Andrew, "Job Queues and Layoffs in Labor Markets with Flexible Wages,"
Journal of Political Economy, LXXXVIII (1980), 529-38.
This content downloaded from 147.251.55.59 on Fri, 06 Apr 2018 11:06:10 UTC
All use subject to http://about.jstor.org/terms