ownership form, managerial incentives, and the intensity of rivalry

OWNERSHIP FORM, MANAGERIAL INCENTIVES,AND THE INTENSITY OF RIVALRY

GOVERT VROOMPurdue University

JAVIER GIMENOINSEAD

This study investigates how differences in ownership form—between franchised andcompany-owned units—affect managerial incentives and competitive pricing in dif-ferent oligopolistic contexts. We argue that chains may restrict decision making incompany-owned units as a commitment device to maintain high prices in concentratedmarkets and found evidence consistent with this argument. We also found that a unit’sownership form affected its rivals’ competitive behavior. Our results indicate thatcompany-owned units’ ability to raise their own and rivals’ prices in highly concen-trated markets led to their higher performance relative to franchised units.

Over the years, competitive strategy research hasexamined the dynamics of competitive interaction,including market entry and exit, price rivalry, andother forms of competitive engagement. This re-search typically depicts firms as unitary actors: it isa firm that makes competitive decisions about thenature and intensity of competitive behavior, withthe objective of optimizing firm performance. Thisview downplays the heterogeneity of internal or-ganization within firms. In most organizations,competitive decisions are delegated to managerswhose incentives may be imperfectly aligned withthe interests of the organization as a whole. In somesituations, firms may even delegate competitive de-cisions to external independent agents, dealers orfranchisees, while their competitors delegate thesesame decisions to managers inside the firm. Forexample, with the growth of multiunit and multi-divisional organizations in many sectors of theeconomy, units operated by owner-managers may

compete head-to-head with units operated byprofessional managers of multiunit organiza-tions. Unfortunately, competitive strategy re-search has not sufficiently explored the compet-itive interaction among firms with diverseorganizational forms.

Delegation of competitive decisions requires par-allel consideration of the design of incentive andcontrol systems. Incentive and control systems de-termine which decisions or tasks are delegated, thelimits of the delegated authority, and the explicitand implicit incentives for decision makers. Insidean organization, managerial incentives may beshaped by compensation schemes, expectations ofcareer progression, span of control, administrativerules, organizational structure and culture, andother organizational characteristics.1 Differences inincentive systems among organizations may lead todifferent competitive behaviors and may affect theoutcomes of competitive interactions. For example,managers who receive incentives based on marketshare objectives may act more aggressively in theirmarket than those with strict financial goals (Gupta& Govindarajan, 1984). Managers may be rewardedfor adhering to a corporate pricing policy or en-couraged to adapt prices to local market conditions.Managers with incentives based on unit perfor-mance rather than firm-level performance may bemore willing to engage in competitive actions that

The authors would like to thank Jonghoon Bae, Mar-jolein Bloemhof, Marco Ceccagnoli, Olivier Chatain,Karel Cool, Yves Doz, Arturs Kalnins, Brian McCann,Francisco Ruiz-Aliseda, Vicente Salas Fumas, Metin Sen-gul, Xavier Vives, Michael Yaziji, Peter Zemsky, partici-pants at the SMS, AOM, CCC, and Harvard Strategy con-ferences, and seminar participants at HEC, the Universityof Amsterdam, Erasmus University, and Tilburg Univer-sity for their helpful comments. We also appreciate thehelp of Source Strategies, Inc., and the Texas Comptrol-ler’s Office for their support with data-related issues. Wewould like to thank Associate Editor Amy Hillman andthe three anonymous reviewers for their valuable feed-back and suggestions, which helped improve the articlesignificantly.

1 As used here, “managerial incentives” rests on theassumption that managers make economically rationaldecisions that maximize their payoffs. These incentivesthus include, but are not limited to, incentive or bonusplans.

� Academy of Management Journal2007, Vol. 50, No. 4, 901–922.

901

Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder’s expresswritten permission. Users may print, download or email articles for individual use only.

cannibalize other revenue sources of their firm(Chandy & Tellis, 1998; Christensen, 1997). Whencompetitive decisions are delegated to externalagents, like franchisees or dealers, the explicit andimplicit incentives in the relationship determinethe influence of a firm over these external agents.Since external agents are residual claimants, theyhave discretion over all actions not explicitly con-tracted with the firm and strong incentives to focuson their own performance.

Some organizational theories, such as the trans-action cost economics, agency, organizational ecol-ogy, and contingency theories, have examined theoutcomes of competition among organizationalforms. Usually proponents of these theories pro-pose adaptation and selection mechanismswhereby more efficient organizational forms areselected by an environment. Yet often these theo-ries have glossed over how organizational formsinteract in oligopolistic competitive contexts. Theargument has been that competitive selection isdetermined by “economizing” rather than “strat-egizing” (Williamson, 1991). When the intensity ofrivalry is exogenously determined, only superiorefficiency determines the competitive success oforganizational forms. When organizational formsinteract in oligopolistic competitive settings, how-ever, the organizational choices of a firm affect itscompetitive actions, the intensity of competitiveinteractions in the market, and, ultimately, its per-formance outcomes. Consideration of the strategiccompetitive effects of organizational forms is im-portant for understanding their ultimate perfor-mance, yet this dimension has been largely ignoredin the existing literature.

This study contributes to the literature by devel-oping a theory of how organizational form can beused as a commitment device to influence compet-itive interaction. Specifically, we show that chainorganizations may use company-owned units aslocal price leaders, effectively attenuating the in-tensity of rivalry. Multiunit organizations, rangingfrom multiunit retail chains to multidivisional ormultinational corporations, are important in theU.S. economy (Greve, 2003) and have considerableshares in some industries, such as the restaurant,retail, hotel, and auto service industries (Ingram &Baum, 1997). Forms of unit ownership may varyover and within these industries. For example, bot-tlers carry out the local competitive activities ofsoft drink companies. Some bottlers are fullyowned by soft drink companies, others are partiallyowned by the companies, and still others are inde-pendently owned. Multinationals can consist offully owned units, joint ventures, and units man-aged by independent foreign agents under contract.

Retail chains may have units that are company-owned and units that are franchised. Here, we fo-cus on this latter case as we examine the effect ofthe ownership form of hotel units on their compet-itive behavior in their local markets.

We tested our hypotheses in data on the Texashotel industry from 1997 to 2002. The panel dataallowed us to examine the relationship betweenownership forms and competitive behavior over awide variety of competitive contexts while control-ling for alternative explanations and sources of un-observed heterogeneity. A better understanding ofthe competitive consequences of ownership formswould allow firms to design organizations that aremore effective at credibly carrying out their com-petitive strategy. Moreover, in contrast with thecurrent emphasis on efficient organizational de-sign, our view would allow firms to use ownershipforms that effectively balance efficiency and strate-gic benefits.

THEORY AND HYPOTHESES

Context

Multiunit organizations, which have been dis-cussed extensively (e.g., Greve, 2003; Greve &Baum, 2001; Ingram & Baum, 1997), are defined asorganizations, such as restaurants, hotels, retailchains, and multinationals, that operate in multiplemarkets through several distinct units. One reasonfor recently enlarged research interest in these or-ganizations may be the realization of their rapidlygrowing importance. Ingram and Baum suggestedthat “chains will eventually come to dominate ev-ery service industry that is characterized by somedirect contact between customers and organiza-tion” (1997: 69). Greve and Baum (2001) pointedout that multiunit organizations are not only prev-alent in service industries but are also frequent inthe manufacturing sector. Another important rea-son for a surge in research interest in multiunitorganizations is the challenge this organizationalform poses for conventional theory. Greve (2003)argued that the multiunit form highlights the short-comings of theories that are based on the premise oforganizations as unitary actors in a unitary environ-ment. Multiunit organizations face an interestinginterplay between cross-unit learning and compet-itive behavior. Moreover, as Greve noted, multiunitorganizations have more strategic choices thanother organizational forms.

The relevant literature on multiunit organiza-tions can broadly be divided in two groups. Thefirst concerns the question of what the benefits anddrawbacks of multiunit organizations are, as com-

902 AugustAcademy of Management Journal

pared with single-unit organizations. In their studyof the Manhattan hotel industry, Ingram and Baum(1997) examined benefits and drawbacks of chainaffiliation, arguing that chain affiliation improves alocal unit’s competitive situation by providing re-sources, reputation, and market power. Ingram(1996) highlighted the tension within multiunit or-ganizations between consistency for the chain as awhole and local adaptation. He suggested that themultiunit organization form could solve the prob-lem of free riding through credible commitment tospecified quality levels.

The second stream of research on multiunit or-ganizations is focused on the choice of ownershipform within a multiunit organization—that is, thechoice between company ownership and franchis-ing of units (e.g., Caves & Murphy, 1976). Thischoice is affected by the differential characteristicsof these two ownership forms regarding the abilityto monitor local managers and the incentive forlocal managers to free ride on brand-name capitaland firm-specific assets (e.g., Bergen, Dutta, &Walker, 1992; Brickley & Dark, 1987; Lafontaine,1992). An argument for the use of franchising isthat the relative ease of obtaining franchisee laborand capital facilitates rapid growth for startingfirms (e.g., Combs, Michael, & Castrogiovanni,2004). In a traditional franchise, the franchisor sellsa (semi)finished good to the franchisee; in “busi-ness format franchising,” on the other hand, thefranchisee obtains the right to use the franchisor’strade name and business plan. In the latter case, thefranchise contract typically specifies the fees thefranchisee should pay, such as an initial fee androyalty and marketing fees.

Ownership Forms and Competitive Behavior

This article examines how ownership might af-fect competitive behavior. Ownership confers twoproperty rights: residual control rights and residualincome rights (Hart, 1995). Ownership forms, de-fined as different structures of property rights, thusdiffer in residual control and income rights.2 Own-ership may affect competitive behavior to the ex-tent that ownership shapes the incentives of man-agers who make competitive decisions. Ownershipcould relate both to vertical relationships (i.e.,regarding upstream suppliers and downstreambuyers) and horizontal relationships (i.e., regard-

ing different units supplying similar goods indifferent local markets or segments). In bothcases, one might wonder if competitive behaviordiffers depending on whether the units areowned separately or jointly.

Economics has studied how vertical ownershipforms affect competitive behavior. For example,Spengler (1950) argued that a vertically integratedmonopolist will charge a lower price than two ver-tically adjacent monopolists. The reason is that thedownstream monopolist purchases the goods at amargin above cost and thus applies a profit marginon top of the margin at which the goods were pur-chased. This procedure results in an accumulationof margins referred to as double marginalization.In the absence of competition, double marginal-ization increases the price for the end consumerto an inefficiently high level. However, doublemarginalization may also occur in oligopolisticsituations in which firms have less-than-perfectmarket power. There has been ample empiricalsupport for this argument. For example, Muris,Scheffman, and Spiller (1992) found that con-sumer prices for goods from systems of indepen-dent bottlers were higher than the prices of thosefrom integrated soft drink bottlers.3

Horizontal ownership may also affect competi-tive behavior. For example, if demand spilloversare present, managerial incentives may differ de-pending on whether only one or multiple units areowned (Lafontaine, 1999). Demand spillovers canbe illustrated by the following examples: A lowprice at one unit may increase demand not only forthat unit but also for other units that belong to thesame chain (positive demand spillover). Con-versely, a low price at one unit may cannibalizedemand at other closely located units that are per-ceived as substitutes (negative demand spillover).In the case of independently owned units, the spill-over would be treated as an externality, but thespillover would be internalized when the sameowner owns both units. In conclusion, both verticaland horizontal ownership may affect the competi-tive behavior of individual units.

Another theoretical mechanism linking owner-ship form and competitive behavior that has not yetreceived sufficient scholarly attention is the effectof ownership form on commitment to particularcompetitive responses. In other words, we explorethe question of whether units with different own-ership forms differ in their ability to commit to

2 The term ownership form has been used in the eco-nomics and management literatures to represent unitswith different structures of property rights (e.g., Hans-mann, Kessler, & McClellan, 2003).

3 For a review of the determinants and effects of ver-tical integration, see Perry (1989).

2007 903Vroom and Gimeno

more or less aggressive competitive behavior. Thisquestion is discussed in the next section.

Strategic Commitment through Ownership Form

In this article, we focus on two prevalent owner-ship forms used by multiunit organizations: com-pany ownership and franchising. We explore howthe differences between these two ownership formsmay affect competitive behavior. Although compet-itive interactions among rivals may happen in mul-tiple dimensions of strategy (e.g., Ferrier, Smith, &Grimm, 1999), here we specifically study the pric-ing behavior of these company-owned and fran-chised units and its effect on competitive pricingby their rivals. Firms compete for customers bygiving them superior “value propositions,” definedas the maximum price a customer is willing to payfor the product attributes minus the price actuallypaid. Firms can improve their value propositionsby increasing product and service offerings, qual-ity, and other dimensions valued by customers.However, the price the firm sets determineswhether the firm or the customer appropriatesvalue. This distinction between value creation andvalue appropriation highlights the strategic impor-tance of price setting in competitive contexts. Asecond reason price is an aspect of competition thatwas particularly well suited for our study is thatprices are more adaptable than other dimensions ofcompetition, such as quality or location. Thisadaptability implies that prices are more respon-sive to changes in the competitive landscape andthus reflect competitive interaction well.

Franchising and company ownership differ fun-damentally on two key dimensions of organization-al design that may affect competitive behavior.These key design dimensions relate to the alloca-tion of control rights to units (delegation of author-ity) and the incentives of unit managers (objectives,control systems, and rewards).

The franchising ownership form contractuallyendows residual control rights to a local unit, thefranchisee. In other words, the franchisee is enti-tled to decide about all aspects of its business ex-cept for those that the contract with the franchisorexplicitly delimits. A franchisor or brand ownermay impose certain obligations aimed at guarantee-ing a consistent level of quality throughout a chain,may specify certain product characteristics, or mayengage in nationwide advertising to influence de-mand. Yet these contractual conditions apply to allunits within the chain. Moreover, contracts cannot

legally specify prices at the unit level.4 Thus, fran-chisees set their own price (Yin & Zajac, 2004).

Whereas franchisees have substantial autonomyto set prices, the managers at the corporate head-quarters of a chain are free to decide what level ofautonomy to give to the managers of local units. Acorporate office may let local managers operate rel-atively freely, so long as they respect the qualitystandards that define the brand, or they may decideto provide strict rules and guidelines detailing notonly operational procedures but also competitiveactions such as pricing. Yin and Zajac studied thegovernance differences between franchising andcompany ownership and concluded that “company-owned stores are much more hierarchical and aresubject to much more control and monitoring fromchain operators than franchised stores” (2004: 368).Specifically, they found that “pricing is determinedby headquarters for company-owned stores” (2004:370).5

A second difference between franchising andcorporate ownership concerns incentive and re-ward systems. In the case of franchising, after pay-ing various fees to the franchisor, the franchisee isthe residual claimant for the remaining income, asituation creating powerful incentives. The pay-ment of franchise fees, typically set as a function ofunit revenues, may lead to a distortion of the fran-chisee’s objective function. Because the franchiseesbear the full unit costs but only capture a portion ofunit revenues, their incentives to cut prices to gen-erate additional volumes and revenues is reduced.In a variety of industrial contexts, this distortion(double marginalization) has led franchised unitsto set higher prices than company-owned units (seeLafontaine and Slade [1997] for a review of thisliterature).6 By contrast, in company ownership the

4 In the United States, the autonomy of franchisees toset prices is well anchored in legal contexts. See Lafon-taine (1999) and Blair and Lafontaine (1999) for moredetails about the legal background of resale price main-tenance in franchising.

5 A detailed comparative analysis of governance andother differences between company ownership and fran-chising can be found in Bradach (1998).

6 This effect holds both in the case of traditional fran-chising, in which the franchisee purchases a (semi)fin-ished good from the franchisor, and in the case of busi-ness format franchising, in which the franchisee pays aroyalty fee for the use of a trademark. In the latter case, itis crucial to note that royalty fees are typically a percent-age of revenues rather than a percentage of profits (Lafon-taine & Shaw, 1999). The payment of a fee that is apercentage of revenues distorts competitive behavior inthe same way as double marginalization does, but the


incentive system may vary from case to case.Managers’ reward systems are typically based onfactors such as revenues, costs, market share, oc-cupancy, and other performance measures. Aspecific reward system could lead to double mar-ginalization, but there is no reason to expect thegeneral occurrence of double marginalization asin the case of franchising.

Before we explore how the above-described dif-ferences in control rights and incentive systemsmay affect the competitive behavior of company-owned and franchised units, we first analyze thechallenges the units of multiunit organizations typ-ically face. These units generally compete in localmarkets in retail or service industries in whichgoods are differentiated. Because competition oc-curs in a limited geographic area, it is oligopolistic,even though the number of local competitors var-ies. Oligopolistic competition generally drivesprices below the monopoly level, which impliesthat (modest) increases of the average market pricewould have a positive effect on industry profitabil-ity. When oligopolistic competition focuses on theprice dimension, firms need to understand thattheir own and rivals’ prices in general move in thesame direction: if a focal firm decreases its price,competitors’ residual demand decreases, and gen-erally it is in the interests of direct competitors todecrease prices too, while the opposite is true ifthe focal firm increases its price (Bulow, Geana-koplos, & Klemperer, 1985). This scenario sug-gests that playing the role of price leader in-creases profits for both a firm and its competitorsif rivals follow (Fudenberg & Tirole, 1984). How-ever, rivals only follow a price leader’s action if theprice commitment is credible (Schelling, 1960). Weargue that the organizational differences betweenfranchised and company-owned units affect theirability to act as price leaders. Specifically, althoughcompany-owned units can credibly commit to sus-taining price discipline in their local market, fran-chised units find it difficult to do so, as explainedbelow.

The value of commitment as a way of strategi-cally influencing one’s competitors has been acrucial topic of research in competitive strategyand game theory (see, for example, Dixit, 1980;Ghemawat, 1991). A commitment that is observ-able, understandable, and credible limits the ac-tion space of an actor, which affects the payoffs ofits rivals and, ultimately, the optimality of itsrivals’ reaction possibilities (Besanko, Dranove, &

Shanley, 2000). A company-owned unit can cred-ibly commit to increasing its price level becauseof the administrative constraints created by thedelegation relationship between the center andlocal unit. The manager of a company-ownedunit must operate within the rules and guidelinesset by corporate headquarters. In other words, amanager of a company-owned unit can crediblycommit to setting relatively high prices becausecorporate headquarters told them to do so. Sim-ilarly, Bradach referred to company ownership as“a military model” (1998: 1).

For rules and guidelines in a corporate owner-ship context to create a credible commitment, theyhave to be difficult to reverse (Dixit & Nalebuff,1991). One may wonder whether administrativedirectives are indeed fully rigid. In other words,can directives be changed in the face of competi-tion and thus nullify the credibility of the commit-ment? Chains typically develop pricing policieswithin which local managers have to operate. Cor-porate headquarters habitually use annual budgetswith detailed information about price levels fordifferent products and the quantity of the output.These budgets are decided upon after some form ofinternal negotiation with the local units. The obli-gation to act according to the described administra-tive system, consisting of a pricing policy and anannual budget, significantly restricts a local man-ager’s autonomy. This restriction cannot easily bechanged because of the administrative costs. More-over, for reputational reasons within a chain, head-quarters have an incentive not to renege on theseinternal contracts. In sum, we conclude that thesystem of internal control that administers the re-lationship between headquarters and company-owned units creates managerial incentives thatmay indeed be difficult to reverse, at least in theshort term.

Clearly, franchised units face a completely differ-ent situation. A franchisee may want to commit tosetting higher prices, yet he or she cannot crediblydo so. Pricing decisions are irrevocably the respon-sibility of the unit. The unit owner can changeprices and offer discounts in response to customerand competitive demands. As the residual claim-ant, the unit owner may be strongly tempted tolower prices and increase sales, especially in situ-ations where fixed costs are high and capacity isnot fully utilized. The autonomy of the franchiseeand the absence of interference by corporate head-quarters undermine the ability to engage in credibleprice leadership.

Given the limitation on franchisees’ ability to cred-ibly commit to refrain from discounting, one maywonder whether franchisors can help franchisees by

payment of a fee based on profits does not (Schmidt,1994).


using royalty fees to create a credible commitment.As discussed before, royalty fees increase pricesthrough double marginalization and can thus po-tentially decrease product-market rivalry. Influenc-ing rivalry through royalty fees, however, is imper-fect for the following reason: An optimal priceincrease reflects local competitive conditions. Spe-cifically, the possibility to increase prices anddampen competition may exist in oligopolisticmarkets, whereas markets that approach perfectcompetition do not provide the same opportunity.Although the optimal level of royalty fees is con-tingent on the local competitive context, royaltyfees are relatively invariant and cannot adjust tolocal competitive circumstances.7

In conclusion, two conditions are necessary toeffectively create price leadership through thechoice of ownership form. The first condition is theability to commit to a price that constitutes anoptimal price leadership position, given specificlocal competitive conditions. The second conditionis that this commitment be credible and not easilyreneged upon once units compete. Company own-ership provides the possibility of satisfying bothconditions, because administrative rules andguidelines can be set taking local competitive cir-cumstances into account and, once set, they bindunit managers. By contrast, franchising satisfiesneither of the two conditions.

This perspective on multiunit competition re-flects a stream of research in economics referred toas strategic delegation theory. Schelling (1960) dis-cussed the use of delegation of authority to an agentas a means to credibly commit to a certain course ofaction. He provided the example of a negotiation inwhich one party sends a delegated agent as op-posed to participating personally. By choosing anagent with incentives that lead the agent to behavepredictably but differently from the principal, arational player (the principal) can credibly committo an action that otherwise would not be credible.Vickers (1985), exploring the implications of theconcept of strategic delegation for the theory of thefirm, pointed out that though owners typically haveprofit-maximizing incentives, the managers whopopulate a firm do not necessarily have such incen-tives. The presence of delegation in firms can allow

them to benefit from strategic effects that wouldotherwise not be possible. In sum, Vickers arguedthat “the separation of ownership and control of afirm may be a good thing for the owners, becausenon-profit-maximizing managers may earn higherprofits than would profit-maximizers” (1985: 139).Fershtman and Judd (1987) and Sklivas (1987) for-malized the idea of strategic delegation in game-theoretical models). This article can thus be seen asa further development and empirical test of strate-gic delegation theory in the context of competitionamong multiunit organizations.8 The next subsec-tions develop hypotheses that explore the effect ofstrategic commitment and royalty fees on pricingbehavior, strategic interaction, and revenues.

We have discussed above two distinct theoreticalmechanisms that may create a price differentialbetween company-owned and franchised units:strategic commitment through delegation and dou-ble marginalization resulting from royalty fees.These two mechanisms may work in opposite di-rections, because double marginalization causesfranchised units to price higher than company-owned units, but strategic commitment throughdelegation causes company-owned units to pricehigher. However, whether franchised units or com-pany-owned units will price higher in any specificsituation depends on the combination of both ef-fects. Prior research has observed that prices ofcompany-owned units on average tend to be lowerthan those of franchised units (Lafontaine & Slade,1997) but has not been able to explicitly tie thedifference to an underlying causal mechanism. Thechallenge we faced was to develop hypotheses thatexplicitly separate these two effects—strategiccommitment and double marginalization—so thateach could be tested independently. Our strategywas to explore contingencies that make each of the

7 Royalty fees are relatively invariant for several rea-sons. First, franchise contracts have a typical duration of15 years, and contract terms are very persistent over time(Lafontaine & Shaw, 1999). Moreover, the typical termsof franchise agreements are public knowledge, whichlimits the possibility of offering different terms to differ-ent franchisees (see, for example, the 2001 Hotel Fran-chise Fees Analysis Guide, from HVS International).

8 Delegation can be defined as the handing over of atask or a decision, usually from one organizational levelto a lower one. The existing strategic delegation literaturehas focused on delegating decision-making authority toagents that have been given incentives different from theprincipal’s. Under those conditions, the combination ofagent’s autonomy and distorted incentives provide acommitment device for the principal. For example, Bon-nano and Vickers (1988) suggested that commitmentcould be accomplished by contractually separating amarketing unit and removing the incentive to cut priceby using a high royalty fee. In our study, commitmentoccurs via use of organizational rules that constrain theagent’s decision range—such as restricting their author-ity over price—rather than via modification ofincentives.


two mechanisms more salient, allowing us to mea-sure the magnitude of each effect separately.

Effect of strategic delegation on competitivepricing. As discussed above, in the case of oligopo-listic competition with differentiated goods, an op-timal incentive system encourages local units tokeep prices relatively high and refrain from dis-counting. Providing such an accommodating incen-tive results in an increased price level for a focalfirm and encourages a similar response from directcompetitors. The benefit of commitment, however,depends on the extent to which strategic interac-tion plays a significant role in a given market. Asthe number of firms in a market increases and mar-ket concentration decreases, strategic interactiondecreases, and the benefit of strategic commitmentdissipates. In a market with a large number of com-petitors, communication and coordination becomemore difficult, which decreases the potential gainof commitment (Stigler, 1964). In the extreme caseof perfect competition, there would be no strategicinterdependence, and firms would not set incen-tives that induce accommodation. Thus, any pricepremium resulting from strategic delegation shouldincrease as a market becomes more concentrated.

The ability to credibly commit to increasingprices allows chains with company-owned units tobenefit from strategic delegation when that wouldpay off most: in highly concentrated markets. Con-versely, because of the full autonomy of franchisedunits regarding their pricing behavior, these unitscannot credibly commit to setting high prices. Thedifference in the governance mechanism betweenfranchised and company-owned units leads to thefollowing hypothesis:

Hypothesis 1. Market concentration has astronger positive effect on company-ownedunits’ prices than on franchised units’ prices.

Effect of royalty fees on competitive pricing.Because of double marginalization, royalty feesprovide the incentive to franchisees to increase theprice charged to the end customer. Schmidt (1994)showed theoretically that this distortion directlydepends on the magnitude of royalty fees. Al-though prior research has shown that pricescharged by franchised units are on average higherthan those charged by company-owned units(Lafontaine & Slade, 1997), no study, to our knowl-edge, has evaluated the magnitude of such a pricedifference and its relation with royalty fees.

Hypothesis 2a. The price charged by fran-chised units is positively associated with therate of the royalty fees these units pay, exhib-iting a “pass-through effect.”

The ability to accumulate margins, however, de-pends on the competitive context. In the extremecase of perfect competition in a final market, inwhich no individual firm has any market power,the market price is given and the cost structure mayaffect survival chances but not pricing behavior.Tirole (1988) posited that the markup would bezero in the case of perfect competition and maximalin the case of monopoly. The extent to which fran-chised firms can pass through royalty fees to cus-tomers thus critically depends on the competitivecontext.9 This argument leads to the followinghypothesis:

Hypothesis 2b. The pass-through of royaltyfees in a retail price is positively associatedwith market concentration.

Effect of ownership form on competitive inter-action. Hypothesis 1 states that company-ownedunits increase prices more than franchised units asconcentration increases. The underlying argumentwas that it is favorable for any firm to increaseprices, but only company-owned units are capableof credibly committing to doing so. However, in-creasing price is only favorable if rivals follow byincreasing their prices too. The direct effect of aprice increase is a quantity decrease, so the effecton revenues, and ultimately profits, is ambiguous.The indirect or strategic effect of a price increase,by contrast, is that rivals increase their price too(Bulow et al., 1985), which increases a focal firm’sdemand and, thus, its revenues. Consequently, weexpect being in direct competition with company-owned units in a concentrated market to increase aunit’s price, which leads to the followinghypothesis:

Hypothesis 3. Market concentration has astronger positive effect on prices of the compa-ny-owned units’ rivals than on prices of thefranchised units’ rivals.

Effect of ownership form on performance. Inthe absence of cost information, managementscholars and industry practitioners frequently userevenues per unit of capacity as a performancemeasure, especially in industries with high fixedcosts, such as the airline and hotel industries (e.g.,Canina, Enz, & Harrison, 2005). In this study, we

9 Note that royalty fees are not the same for all marketparticipants; independent and company-owned units donot pay royalty fees, and the level of royalty fees differsper brand. Although industry-wide cost fluctuations maybe passed through to customers even in the case of per-fect competition, firm-specific costs would not be.


followed that practice. Hypothesis 1 asserts thatcompany-owned units set higher prices than fran-chised units in concentrated markets. In line withthis, Hypothesis 3 argues that in concentrated mar-kets company-owned units’ rivals set relativelyhigh prices. Hypotheses 1 and 3 taken togethersuggest that company-owned units play the role oflocal price leaders, that rivals follow their priceincreases, and thus that company-owned units ef-fectively lessen rivalry, thus raising performance.This argument leads to the following hypothesis:

Hypothesis 4. As market concentration in-creases, the revenues per unit of capacity ofcompany-owned units increase more thanthose of franchised units.

METHODS

Sample

The empirical setting of this study was the hotelindustry, which is characterized by local competi-tion: hotels compete directly with establishmentslocated in their same area but not with hotels lo-cated in a completely different part of their countryor state (Baum & Mezias, 1992). Hotels compete onlocation choice, quality, range of product offerings,and, after these are determined, price. This studyfocused on price, which reflects the short-run com-petitive interaction between hotel establishments.As discussed in the previous section, price consti-tutes an important aspect of competition, highlight-ing the difference between value creation and ap-propriation, and prices are well suited to thisstudy’s aims because they are more responsive tochanges in the competitive landscape and thus re-flect short-run competitive interaction well. More-over, prices are easily quantifiable, which facili-tates statistical analysis, and are accessible toresearchers. Finally, the relatively sticky nature ofother factors that influence pricing, such as thequality of product offerings, allows for effectivecontrol when using panel data.

Several studies have examined the benefits anddrawbacks of horizontal ownership forms withinchain organizations in the hotel industry. For ex-ample, Ingram and Baum (1997), Chung andKalnins (2001), and Canina, Enz, and Harrison(2005) explored agglomeration spillovers in hotelchains. Ingram (1996) argued that chain organiza-tion can solve the problem of commitment to qual-ity. Other studies have explored the difference inbehavior between company-owned and franchisedunits, particularly focusing on the internalizationof competitive spillovers within a chain. Conlin(2002), for example, showed that within a given

chain, competition between franchised units isfiercer than it is between company-owned units,because franchisees do not consider how their pric-ing behavior affects demand for other units. In astudy of intrachain competition, Kalnins (2004)showed that the entry of franchised units into amarket decreases the revenues of the incumbenthotels of the same chain, but the entry of company-owned units increases revenues of incumbents.These studies, however, do not explore the effect ofownership form on pricing across different chains.

The data we used were from the Texas hotelindustry. Because of the large size of the Texasmarket, the heterogeneity of the state in terms ofrural and urban regions, and the availability ofreliable data, the Texas hotel market is well suitedfor empirical research. The two primary sources ofdata we used are a hotel tax file we obtained fromthe State of Texas Comptroller’s Office and theSource Strategy, Inc., database. Chung and Kalnins(2001), among others, used the Comptroller’s data-base, and Conlin and Kadiyali (2006) and Conlin(2002) used the Source Strategy database.

All hotels in Texas are required to report reve-nues quarterly to the State of Texas Comptroller’sOffice, which makes these data publicly available.This data set included the hotel name, the locationof the hotel, the name and address of the owner, thenumber of rooms available, and the quarterly reve-nues. The Source Strategy, Inc., database reportedwith the same periodicity and included the samehotels, except those with room revenues below$16,000 per quarter. The data comprised the nameof the hotel, the brand name if the hotel belonged toa chain, its location (town and zip code), the aver-age quarterly occupancy rate, price, and revenueper available room.10 The complete data set, whichspanned 24 quarters (1997–2002) and contained76,906 observations, provided unique research op-portunities because it contained fine-grained longi-tudinal competition data as well as detailed data atthe establishment level.

Our baseline specification defined the local mar-ket as the zip code area in which a focal hotel issituated, in line with Chung and Kalnins (2001). Analternative, firm-centric market definition based ondistance to a focal unit was used for robustnesschecks and is discussed in the Discussion and

10 The average room price (average daily rate), theoccupancy rate, and the average revenue per availableroom are the three most commonly used performanceindicators in the hotel industry. The relationship be-tween these three measures is as follows: revenue peravailable room � occupancy � average room price.


Conclusion section. In Texas, approximately3,500 hotel establishments compete in more than800 local markets. During the period of our sam-ple, 1997–2002, the number of hotels in Texasgrew by 5 percent annually. Franchising grewfrom 30 to 34 percent by the end of 2002. Thenumber of company-owned units grew slightly,from 13 to 15 percent. Independent hotels formed57 percent in 1997 and 51 percent in 2002. Thenumber of hotels per zip code in our sampleranged from 1 to 54. The number of markets inwhich only one establishment operated (monop-oly markets) equaled 31 percent. In 61 percent ofthe markets, between 2 and 10 hotels operated. In8 percent of the markets, more than 10 propertieswere present.

Although our measures of market characteristics,such as average occupancy rate and market concen-tration, included all hotels in the market, the re-gressions only included observations of brandedchains (independent hotels were excluded), reduc-ing our sample to 36,668 observations. Comparingchains, consisting of company-owned and fran-chised hotels, with independent hotels is difficultbecause of the importance of the brand reputationsthat the chains have built. Moreover, the data onquality available for chains were not available forindependent hotels, which would have made iteven more difficult to compare chains withindependents.11

We used brand dummies rather than segmentdummies because within-segment, across-branddifferences are significant in the hotel industry. Forexample, a large business hotel such as the Westinis in many respects different from a niche hotelsuch as the Four Seasons, even if both are in theluxury segment. Each brand belongs to only onesegment.

Some brands chose to have only company-ownedor franchised units. In these cases, it was impossi-ble to distinguish between the brand effect and theownership effect. Of the 82 hotel brands that werein business in Texas, 28 brands operated both fran-chised and company-owned hotels. We defined abrand as belonging to this category if at least 5percent of its units were franchised and at least 5percent were company-owned. Limiting our analy-sis to brands that contained both franchised andcompany-owned units reduced our sample to12,069 observations. Controlling for brand and lim-

iting our focus to those that had both franchisedand company-owned units allowed us to identifythe effect of ownership form within a commonbrand. Table 1 contains a list of the includedbrands, including information about the marketsegment, the average price per room, the percent-ages company-owned and franchised, the royaltyfee, and the number of units in Texas.

We treated the choice of ownership form as ex-ogenous to our model. As discussed in the Theorysection, a considerable literature examining thechoice of corporate ownership versus franchisingexplores several theoretical perspectives, includingagency theory and imperfect capital and labor mar-kets. Theoretically, it is possible that hotel corpo-rations take the strategic commitment effect ofownership choice and its potential performanceconsequences, as explored here, into account whenmaking the ownership form decision. However,given the novelty of these ideas and the salience ofother theoretical reasons to choose between corpo-rate and franchisee ownership, we do not considerthis a significant problem. In the section on controlvariables, we discuss how we dealt empiricallywith possible selection effects of ownership form.

Dependent Variables

Price per room per night. The primary depen-dent variable was the average price for a hotelroom. The relative variability and easy measurabil-ity made price an empirically powerful aspect ofcompetition, especially because the panel data setallowed effective control for the relatively sticky,nonquantifiable aspects of competition, such asquality and location. We used the natural logarithmof price, because price followed an approximatelognormal distribution. Given controls for other es-tablishment, brand, and market attributes, lowerprices reflect more intense rivalry. The price mea-sure was an average for each hotel establishmentfor each period. In reality, the price per night in ahotel room usually fluctuates according to thelength of the stay, the size of the room, the day ofthe week, and the season. Because our variable wasaveraged per three-month period, it did not capturethese price differences. We do not believe that thisdata limitation systematically biased our findings.

Rivals’ price per room per night. A second de-pendent variable was the average price the rivalhotels in the same market charged. This variablewas calculated by dividing the sum of rivals’ quar-terly revenues in a specific market by the totalnumber of room nights sold by rivals in the sameperiod and market. As described above, we used

11 Some hotel tour books, such as the one published bythe American Automobile Association, report a “star rat-ing,” but these cover at the most 15 percent of the hotelsin Texas (Chung & Kalnins, 2001).


both a zip code and a firm-centric marketdefinition.

Revenues per available room. Revenues peravailable room (typically abbreviated as “RevPAR”by industry practitioners) was used as our perfor-mance measure. We calculated this variable by di-viding the total quarterly revenues by the numberof rooms that were available at a focal hotel.

Independent Variables

Ownership form. The main independent vari-able was ownership form. We created a dummyvariable for franchised establishments (1, “fran-chised,” 0, “not franchised”). The company-ownedownership form was the omitted variable. We de-termined whether a unit was franchised or compa-

ny-owned in a certain quarter using the brandname, the owner address, and information fromthe Directory of Hotel & Lodging Companies(2003).12, 13

12 The Directory of Hotel & Lodging Companies pro-vides the addresses (town and state) of corporate head-quarters. All hotels located in the same town as theirowners are assumed to be company-owned. All otherchain hotels are assumed to be franchised.

13 Some hotels are “under management contract,”which can be considered a separate ownership form. Inthe case of management contracts, an outside investorgroup typically owns the hotel and outsources its man-agement to a franchisor. We treated hotels in this cate-gory as if they were company-owned because “the man-agers of these hotels . . . face the same incentive problems

TABLE 1Brands Operating Both Franchised and Company-Owned Hotels in Texas

Segment and Brand

AveragePrice in

U.S. DollarsOccupancy

RateRoom

CapacityRoyalty

FeePercent

Franchised

PercentCompany-

OwnedNumber of

Hotels

EconomyCountry Hearth Inn 50.55 55.4% 42 4.00% 28% 72% 8Motel 6 36.95 64.7 110 4.00 14 86 95Red Roof Inns 44.62 60.1 129 4.48 24 76 28Shoney’s Inns & Suites 50.56 55.0 98 3.50 33 67 7Studio 6 35.69 66.3 131 5.00 17 83 6

Midscale without food and beverageBaymont Inns & Suites 49.58 57.2 105 5.00 21 79 10Candlewood Hotels 59.52 63.5 121 4.75 13 87 11Hampton Inns & Suites 66.68 66.0 105 4.00 91 9 61Mainstay Suites 61.59 55.0 80 4.50 72 28 4Sleep Inns 53.32 55.9 84 4.50 60 40 9Town Place Suites 68.66 60.1 106 5.00 62 38 5Wellesley Inn & Suites 53.64 56.4 124 4.50 15 85 7

Midscale with food and beverageClub Hotel by Doubletree 72.41 59.0 190 3.88 61 39 4Harvey Hotels 82.33 61.2 314 4.52 31 69 5Holiday Inn 68.58 61.2 214 5.04 83 17 78

UpscaleAmeriSuites 76.82 63.3 127 5.00 22 78 15Courtyard by Marriott 87.86 70.0 137 5.42 58 42 37Crowne Plaza Hotels 76.85 62.4 278 5.00 37 63 6Homewood Suites by Hilton 100.53 65.5 103 4.00 69 31 16Radisson 79.72 59.2 272 4.00 87 13 14Residence Inn by Marriott 98.95 72.6 110 4.96 59 41 29Staybridge 88.48 60.2 111 5.00 20 80 3Wyndham Hotels & Resorts 110.05 61.0 404 5.00 36 65 8

Upper upscaleDoubletree Suites 131.99 73.0 269 3.94 54 46 2Doubletree 108.02 67.7 374 3.94 86 14 6Embassy Suites 113.12 69.2 224 4.00 45 55 16Omni Hotels 112.87 64.0 363 3.00 11 89 9Sheraton Hotels/Inn/Resorts 97.01 61.6 308 5.00 73 27 10


Market concentration. The most commonlyused measure of market concentration is the Her-findahl index, which is defined as the sum of thesquared shares of the firms in a given market. Wecalculated the Herfindahl index on the basis of thecapacity share (number of rooms available) of allfirms in a market.14

Royalty fees. Hotel & Motel Magazine annuallypublished a “Franchising Supplement” over the1997–2002 study period, with detailed informationabout the franchising fees per brand. Our variableindicated the royalty fee per brand as a percentageof revenues. Franchise fees were relatively stableduring the period of observation (Lafontaine &Shaw, 1999). Moreover, changes in the fees werereflected only in the new contracts that were signedafter the changes took place. We therefore includedaverage franchise fees per brand over the six-yeartime period.

Control variables. Because our data were a panelof multiple hotel establishments from multiplechains observed in multiple markets at multipleperiods, we used fixed effects to control for sourcesof heterogeneity that we could not explicitly mea-sure. These fixed effects also account for possiblenonindependence of observations that share thesame sources of variance. First, we includeddummy variables for each period (year and quarter)to control for unobserved seasonal and macroeco-nomic trends in the Texas economy. Second, weincluded fixed-effect dummies for the brands in thesample. These dummies controlled for the averagequality, reputation, and hotel profile of hotelswithin a given brand.15 Our use of brand dummiesprovided a superior control for unobserved differ-ences than previously used quality segment dum-mies. Brand dummies also control for possible non-independence of observations resulting fromcoordinated pricing strategies within brands. Thebrand dummy was based on the brand name foreach establishment/period combination as reportedin the Strategy Source Inc. database. We cross-checked the brand code with the hotel name re-ported by the Comptroller’s Office. Finally, we in-cluded fixed-effect dummies for each physical

establishment. These dummies accounted for thedifferences in hotel location and stable hotel char-acteristics, such as hotel size and the quality of theinfrastructure.

Establishment dummies were also helpful to ac-count for potential endogeneity of ownership form.It is possible that chains base ownership formchoices on characteristics of an establishment orlocation. If that is the case, estimates of ownershipform might suffer from self-selection bias (Hamil-ton & Nickerson, 2003), in that unobservable char-acteristics relevant in the choice of ownership formalso influenced the competitive behavior of the es-tablishment. Statistically, that bias would be re-flected by a correlation between unobservable es-tablishment characteristics and the ownership formvariable. The self-selection problem can be ad-dressed by examining the Hausman test associatedwith fixed establishment effects because the Haus-man statistic explicitly tests whether observed vari-ables are correlated with unobserved sources ofheterogeneity.

In addition to the above fixed effects, we con-trolled for time-varying conditions of supply anddemand affecting each particular establishment. Tocontrol for supply variations, we included the log-arithm of market capacity, defined as the total num-ber of rooms of all establishments in a market. Tocontrol for market-specific demand shocks, we in-cluded the mean occupancy rate in a local market,defined as the percentage of occupied rooms in theestablishments in the market, excluding the focalfirm. This exclusion was necessary to avoid poten-tial endogeneity because the firm’s own occupancymight be influenced by its price.

Estimation

Our primary model, a reduced-form price equa-tion for each establishment, was estimated using apanel regression. The functional form can bewritten as:

log( price)it �

Xit � � � Zit � � � �i � �c(i) � �t � it,

where Xit represents the independent variables ofinterest; Zit, the control variables; �i, the establish-ment fixed effects; �c(i), the brand effects; and �t, theperiod effects. Apart from the above-describedmodel, we also used the log of rivals’ price andrevenues per available room as dependent vari-

that are encountered by the managers of company-ownedunits” (Kehoe, 1996: 1486).

14 We used the log of the number of hotels in themarket as an alternative measure of market concentrationto test for robustness. This alternative definition yieldedqualitatively very similar results.

15 Since hotel corporations typically own multiplebrands, the brand dummies distinguish among such dif-ferent brands owned by a single corporation.


ables. The functional form was the same as for theprice equation.16

RESULTS

The final sample constituted an unbalancedpanel of 630 establishments (either company-owned or franchised) in 260 markets observed overa maximum of 24 periods, totaling 12,069 observa-tions. There were 6,106 observations of franchisedhotels and 5,963 observations of company-ownedhotels. We compared the variables for these groupsusing univariate t-tests between the two groups.The univariate comparisons showed that fran-chised hotels tended to have higher prices andrevenue per available room. Franchised hotels op-erated in markets with slightly lower average ca-pacity and in markets where their rivals tended tohave higher mean occupancy, but the ownershipforms did not differ in the average concentration intheir markets. Table 2 presents descriptive statis-tics, including means and both the overall and thewithin-establishment standard deviations and cor-relations. The Pearson correlations confirmed thatfranchised establishments charged higher prices onaverage. Price was positively correlated with mar-ket capacity and market occupancy and negativelycorrelated with market concentration. However,these bivariate correlations are not particularly in-dicative of actual causal effects, given the signifi-cant correlations between multiple variables andthe effects of unobserved heterogeneity.

Table 3 presents the results of our analysis. Toavoid multicollinearity between main effects andinteraction effects, and to facilitate interpretation,we centered the interacting variables (market con-centration and royalty fee) on their respectivemeans (Aiken & West, 1991). Because ownershipform is a policy variable, we were concerned that itwould systematically correlate with characteristicsof a chain, market, or establishment. Accordingly,before testing our hypotheses, we examinedwhether the effect of ownership form on pricingbehavior was robust to the inclusion of severalfixed effects. Models 1, 2, and 3 examine the effectof ownership form on pricing after we controlledfor relevant observed and unobserved effects. Theeffect of franchise form on price was positive and

significant when we were only controlling for mar-ket occupancy, market supply, and period effects.Yet the coefficient turned negative and significantwhen we controlled for brand fixed effects. Thisresult indicates that some hotel chains that weremore prone to franchising were also more likely tocharge higher prices at both franchised and compa-ny-owned units. Examination of Table 1 corrobo-rates that large chains in the economy segmentwere more likely to rely on company-owned estab-lishments (e.g., Motel 6, Red Roof Inn, Studio 6),and some large chains in the upscale and luxurysegments were more likely to use franchised estab-lishments (e.g., Doubletree, Sheraton).

Model 3 includes fixed effects for establish-ments, and the relationship between franchise formand pricing remained negative but was only weaklysignificant. Longitudinal changes in ownershipform for a specific establishment were a necessarycondition to ensure statistical power for the estima-tion of ownership effects in the model with estab-lishment fixed effects. Over the six years of obser-vation, 15.8 percent of the establishments in ourfinal sample changed ownership form, whichmeans that they moved from being franchised tobeing company-owned or vice versa. Therefore, weinferred that the marginal statistical significancewas not a consequence of problems of statisticalpower. The results suggest that the strongly signif-icant negative effects found in model 2 might be aresult of unobserved differences between proper-ties. It appears that within a given chain, franchisedunits were associated with unobserved establish-ment characteristics (mainly characteristics of aproperty itself or its surrounding market) that ex-plained their lower prices. Chains might retain ascompany-owned establishments either the moreexpensive properties or those properties with themost attractive market locations. Industry execu-tives confirmed this interpretation, reporting thatchains prefer to own the larger, more luxuriousproperties of a brand because of the complexity ofmanaging these properties (see also Yeap, 2006).The rejection of the null hypothesis (p � .001) inthe Hausman test for establishment-specific fixedeffects confirmed that sources of unobserved heter-ogeneity were correlated with the observed vari-ables in the model. We interpreted these results asconsistent with self-selection in the choice of own-ership form. The inclusion of fixed effects andmarket-level control variables should alleviateany bias that might be due to self-selection result-ing from time-invariant unobserved characteristicsor time-variant market characteristics (Hamilton &

16 Instead of estimating the three dependent variablesseparately, we could have used simultaneous equationestimation. However, separate estimation should notyield biased estimates in a recursive system such as theone studied here. We therefore chose to use the simplerapproach of separate estimation.


TA

BL

E2

Des

crip

tive

Sta

tist

ics

and

Cor

rela

tion

sa

Var

iabl

esb

Mea

ns.

d.

Min

imu

mM

axim

um

Cor

rela

tion

s

Ove

rall

Com

pan

y-O

wn

edF

ran

chis

ed

Sig

nif

ican

ceof

Dif

fere

nce

Ove

rall

Wit

hin

12

34

56

78

1.P

rice

4.14

4.03

4.25

p�

.001

0.41

0.08

3.01

5.33

.42

.57

.00

�.0

9.0

1.2

0.0

82.

Riv

als’

pri

ce4.

144.

144.

140.

360.

092.

305.

32.5

6.2

7.0

1.0

0.0

0.2

1�

.05

3.R

evP

AR

3.68

3.57

3.79

p�

.001

0.50

0.22

0.69

5.00

.89

.49

.02

.00

.02

.36

�.0

54.

Fra

nch

ise

own

ersh

ip0.

510.

001.

00n

.a.d

0.50

0.15

0.00

1.00

.27

.01

.22

.00

.99

.04

�.0

15.

Mar

ket

con

cen

trat

ion

0.21

0.22

0.21

p�

.001

0.17

0.05

0.04

1.00

�.1

9�

.25

�.1

7�

.01

.00

�.2

8�

.71

6.R

oyal

tyfe

ec2.

310.

004.

56n

.s.

2.31

0.68

0.00

5.42

.28

.02

.22

.99

�.0

2.0

3�

.01

7.M

ean

mar

ket

occu

pan

cy0.

580.

580.

59p

�.0

010.

130.

080.

000.

90.2

5.3

5.3

4.0

3�

.49

.04

�.0

38.

Mar

ket

cap

acit

y6.

876.

906.

85p

�.0

10.

910.

143.

569.

00.4

4.5

9.3

7�

.03

�.7

5�

.01

.35

aA

llco

rrel

atio

ns

wit

han

abso

lute

valu

ela

rger

than

.02

are

sign

ific

ant

atp

�.0

5.C

orre

lati

ons

are

wit

hin

the

lim

ited

sam

ple

of28

bran

ds

that

hav

ebo

thfr

anch

ised

and

com

pan

y-ow

ned

un

its

(n�

12,0

69).

Ital

iciz

edco

rrel

atio

ns

are

wit

hin

-est

abli

shm

ent;

the

oth

eren

trie

sar

eze

ro-o

rder

Pea

rson

corr

elat

ion

s.T

he

valu

esfo

rth

ew

ith

inst

and

ard

dev

iati

ons

and

corr

elat

ion

sw

ere

calc

ula

ted

afte

rtr

ansf

orm

atio

nof

each

vari

able

toth

ed

evia

tion

rela

tive

toth

ees

tabl

ish

men

tmea

n.T

his

pro

ced

ure

,kn

own

asab

sorp

tion

,all

ows

for

exam

inat

ion

ofth

est

and

ard

dev

iati

onan

dco

rrel

atio

nof

vari

able

sw

ith

ina

par

ticu

lar

esta

blis

hm

ent

over

tim

e.b

Th

ep

rice

vari

able

s,R

evP

AR

(rev

enu

esp

erav

aila

ble

room

),an

dm

arke

tca

pac

ity

are

loga

rith

ms.

cB

yd

efin

itio

n,

the

roya

lty

fee

equ

als

zero

for

com

pan

y-ow

ned

un

its.

dB

yd

efin

itio

n,

ther

eis

no

vari

ance

inth

eow

ner

ship

form

du

mm

yw

ith

inea

chca

tego

ry.

TA

BL

E3

Res

ult

sof

Reg

ress

ion

An

alys

esa

Ind

epen

den

tV

aria

bles

Fu

llS

amp

leP

arti

alS

amp

les

Mod

el1:

Pri

ceM

odel

2:P

rice

Mod

el3:

Pri

ceM

odel

4:P

rice

Mod

el5:

Pri

ceM

odel

6:P

rice

Mod

el7:

Pri

ce—

Mon

opol

yH

otel

sE

xclu

ded

Mod

el8:

Riv

als’

Pri

ce—

Mon

opol

yH

otel

sE

xclu

ded

Mod

el9:

Rev

PA

R—

Mon

opol

yH

otel

sE

xclu

ded

Fra

nch

ised

own

ersh

ip0.

22**

*(.

01)

�0.

02**

*(.

00)

�0.

01†

(.00

)�

0.01

(.00

)0.

00(.

00)

0.00

(.00

)0.

00(.

00)

0.00

(.01

)0.

01(.

01)

Mar

ket

con

cen

trat

ion

0.20

***

(.02

)0.

19**

*(.

02)

0.18

***

(.02

)0.

12**

*(.

03)

0.14

***

(.03

)0.

60**

*(.

08)

Fra

nch

ised

own

ersh

ip�

mar

ket

con

cen

trat

ion

�0.

10**

*(.

02)

�0.

08**

*(.

02)

�0.

07**

(.02

)�

0.12

***

(.03

)�

0.08

**(.

03)

�0.

23**

*(.

07)

Fra

nch

ised

own

ersh

ip�

roya

lty

fee

0.03

***

(.01

)0.

03**

*(.

01)

0.03

**(.

01)

�0.

04**

*(.

01)

0.05

*(.

02)

Fra

nch

ised

own

ersh

ip�

roya

lty

fee

�m

arke

tco

nce

ntr

atio

n

0.12

***

(.03

)0.

09*

(.04

)0.

13**

(.05

)0.

60**

*(.

10)

Mea

nm

arke

toc

cup

ancy

0.29

***

(.03

)0.

12**

*(.

01)

0.15

***

(.01

)0.

19**

*(.

01)

0.19

***

(.01

)0.

19**

*(.

01)

0.22

***

(.01

)0.

23**

*(.

01)

1.09

***

(.03

)M

arke

tca

pac

ity

0.18

***

(.00

)0.

07**

*(.

00)

�0.

05**

*(.

01)

�0.

01†

(.01

)�

0.01

*(.

01)

�0.

02*

(.01

)�

0.02

**(.

01)

�0.

07**

*(.

01)

0.03

(.02

)

Per

iod

fixe

def

fect

sY

esY

esY

esY

esY

esY

esY

esY

esY

esB

ran

dfi

xed

effe

cts

No

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Est

abli

shm

ent

fixe

def

fect

sN

oN

oY

esY

esY

esY

esY

esY

esY

es

Obs

erva

tion

s12

,069

12,0

6912

,069

12,0

6912

,069

12,0

6911

,808

11,8

0111

,808

Est

abli

shm

ents

630

630

630

630

630

630

619

618

619

R2

.28

.84

.97

.97

.97

.97

.97

.95

.86

R2

(wit

hin

).2

5.2

5.2

5.2

5.2

6.1

6.2

4F

184*

**1,

168*

**10

8***

105*

**10

2***

100*

**10

2***

53**

*91

***

Lag

ran

gem

ult

ipli

erte

st76

,856

***

74,5

09**

*74

,257

***

74,1

10**

*72

,063

***

79,6

50**

*27

,703

***

Hau

sman

test

for

esta

blis

hm

ent

fixe

def

fect

s

3,54

2***

3,25

6***

63**

6,36

3***

3,90

8***

937*

**12

7***

aM

arke

tis

def

ined

byzi

pco

de

inal

lth

ese

mod

els.

Pri

ce,

revP

AR

(rev

enu

esp

erav

aila

ble

room

),an

dm

arke

tca

pac

ity

are

loga

rith

ms.

Sta

nd

ard

erro

rsar

ein

par

enth

eses

.†

p�

.10

*p

�.0

5**

p�

.01

***

p�

.001

Nickerson, 2003). All subsequent models includeperiod, brand, and establishment fixed effects.17

Models 4 to 9 include the tests of the hypotheses.Hypothesis 1 suggests that company-owned unitsincrease their prices more than franchised unitswhen market concentration increases. Model 4 in-cludes the effects of market concentration and theinteraction of market concentration with franchiseform. Because company-owned properties were theomitted category, the results indicated a positiveeffect of market concentration on log-prices of com-pany-owned facilities (slope: .20, p � .001). Forfranchised units, the effect was determined by thesum of the coefficients (slope: .20 � (–.10) � .10,p � .001). The interaction term represented thedifference between the concentration-price slopesfor company-owned and franchised establish-ments. The difference is negative and statisticallysignificant (–.10, p � .001), in support of Hypoth-esis 1. The evidence suggests that company-ownedunits were more likely to raise prices than fran-chised units when market concentration increasedand strategic interactions became most salient.18

Similar results were obtained when a one-periodlagged price was added to the regression as anindependent variable.19 Figure 1a depicts the rela-tionship between concentration and price for com-pany-owned and franchised units.

Hypothesis 2a suggests that higher royalty fees tofranchised units are passed on to customers in theform of higher prices, per double marginalizationarguments. Model 5 adds a variable for royalties.The variable was centered on the mean of 4.56percent for franchised units and was set to 0 forcompany-owned establishments. The coefficienton the royalty fee was equal to .03 (p � .001), which

suggests that a one-point increase in royalty is as-sociated with a 3.1 percent increase in the pricecharged to customers.20 The royalty fees rangedfrom 3 to 5.4 percent in the sample (1.56 below themean to 0.86 above the mean), and the predictedprice effects therefore ranged from 4.7 percent be-low the mean to 2.7 percent above. The resultsconfirm Hypothesis 2a. The insignificance of thefranchise dummy suggests that company-ownedand franchised units are not found to price sig-nificantly differently after the effect of royaltyfees and the differential commitment ability arecontrolled for.

Hypothesis 2b suggests that the franchisor’s abil-ity to pass on increases of royalty fees to customersis limited in more competitive markets. The posi-tive interaction in model 6 between the royalty feeand market concentration (.11, p � .001) confirmsthe hypothesis. The sensitivity of prices to royaltyfees is plotted in Figure 1b for firms at differentconcentration levels. As a market becomes morecompetitive, passing the royalty fee on to custom-ers becomes less feasible. In markets with higherconcentration, however, higher royalty fees resultin even higher customer prices.

In model 7, we examined whether the differentialprice effect of company-owned and franchisedunits was due to monopolistic or oligopolistic pric-ing. Theory suggests that the credibility of strategicincentives should be particularly important in oli-gopolistic contexts because in these situations thestrategic interaction is most salient. Because of thedelegation of pricing to professional managers,company-owned units are able to credibly committo price leadership, and franchised units are not.Since raising prices in monopoly situations doesnot involve competitive interactions, it does notrequire competitive credibility. Consequently, bothfranchised and company-owned units should beable to raise prices to monopoly levels equally well.Model 7 presents the results for the analysis usinga restricted sample that eliminates cases in whichan establishment was the only hotel in a zip codemarket. The predicted slopes of market concentra-tion on the log of price were .12 and .00 for com-pany-owned and franchised units, respectively.With monopoly markets eliminated, the slopes ofboth ownership forms are smaller than in model 6,as would be expected in view of the restriction ofrange in the independent variable. However, thedifference of the slopes is now greater than inmodel 6, and the slope of the franchised units be-

17 Exclusion of brand effects while keeping the estab-lishment effects yielded very similar results.

18 In an auxiliary analysis, we identified how changesin concentration and ownership form, respectively, con-tributed to the results. Clearly, given our fixed-effectsmodel, results would be insignificant without some vari-ation in ownership form and concentration. We carriedout two separate analyses, one focused on the effect ofconcentration changes for establishment that did notchange ownership form and another focusing on the ef-fect of ownership changes. These analyses showed thateach source of variance was present and contributed tothe results in the predicted way.

19 Because this study does not focus on dynamic ad-justment processes, we feel that it is more appropriate touse a traditional regression approach on levels withfixed-effects controls for unobserved heterogeneity ratherthan to model dynamic changes through the inclusion ofa lagged dependent variable (Greve & Goldeng, 2004).

20 This percentage is calculated as follows:exp(.0313) – 1 � 3.1%.


comes insignificant. This result underscores theinability of franchised units to raise prices underoligopolistic conditions.

Hypothesis 3 explores the effect of the ownershipform of a focal firm on the pricing behavior of rival

firms. In our Theory section, we argue that compa-ny-owned units increase prices in concentratedmarkets because of their intended rivals’ reaction:following the price leader. If this reasoning is cor-rect, the rivals of company-owned units would be

FIGURE 1Graphs of Interactions


observed to increase prices more than the rivals offranchised units as concentration increases. Theresults of model 8 confirm this hypothesis becausethe coefficient of concentration on rivals’ prices forfranchised units is significantly less than that forcompany-owned units (.06 versus .14; the differ-ence is significant at the .01 level). These resultsindicate that the rivals of both company-owned andfranchised units increased prices as concentrationincreased, but the rivals of company-owned unitsincreased their prices more than the rivals of fran-chised units, as expected.

Finally, Hypothesis 4 discusses the effect of own-ership form on performance. Our findings for Hy-pothesis 1 imply that company-owned units areable to raise prices more than franchised units asconcentration increases. This finding, however,does not necessarily imply that company-ownedunits’ performance increase, because the direct ef-fect of a price increase is a trade-off between marginand occupancy volume. However, if a company-owned unit’s price increase is followed by priceincreases by its rivals, as our findings in model 8suggest, the overall intensity of rivalry will de-crease, and the focal firm’s performance will in-crease. This reasoning suggests that revenues ofcompany-owned units increase more than those offranchised units as concentration increases. Model9 shows the results of an analysis in which revenueper available room is the dependent variable in asample that excludes monopoly observations. Theresult indicates that company-owned units ob-tained higher revenues per available room in moreconcentrated markets than franchised units, whichis consistent with the combination of our two pre-vious findings: not only company-owned units butalso their rivals set relatively higher prices as con-centration increases. Thus, franchised units tosome extent benefited from concentration in-creases, yet company-owned units were able tobenefit more.21

DISCUSSION AND CONCLUSION

Although traditional research in competitivestrategy treats each firm as if it consists of a unitary

actor, in this study we explore how the organiza-tion of a firm affects the way it competes vis-a-visits competitors. We have developed arguments thatlink ownership forms with managerial incentivesand competitive behavior. The results are generallysupportiveofourhypotheses.Wefindthatcompany-owned hotels are better able to adjust their pricingto the competitive conditions surrounding themand may serve as price leaders in more concen-trated markets. Strategic delegation theory suggeststhat such price leadership behavior is effective inencouraging competitors to follow with higherprices. Influencing competitive decision making af-ter delegation to franchisees is more difficult be-cause chains do not have direct control over theend consumer price, as franchisors are prohibitedby antitrust regulation from directly influencingfranchisees’ competitive decisions. Royalty feescould be used to indirectly affect pricing, but fran-chisors’ ability to use royalty fees as a competitivecommitment mechanism is shown to be limitedbecause these fees cannot be adjusted to local com-petitive circumstances.

Market definitions are always to some degreearbitrary, because perfect definition is impossible.We therefore checked the robustness of our find-ings by trying alternative geographical and prod-uct-market definitions. Table 4 presents the resultsof these robustness checks. First, we investigatedthe robustness of our geographical market defini-tion by using a distance-based definition of localmarket instead of the zip code definition. We “geo-coded” the address of each hotel using Web-basedEZLocate geocoding services. We then defined therelevant market by the establishments within a ra-dius of two miles from the focal unit. The averagenumber of establishments within a two-mile radiusmarket was approximately equal to that within azip code market. The results with this market def-inition (model 10) were largely consistent with pre-vious results (model 7). The difference of slopebetween ownership forms remained significant,with an increased magnitude on the slope of con-centration for company-owned units. The royaltyfee maintained a significantly positive effect onprice, which increased with concentration, asexpected.22

A second robustness check on the market defini-tion took into account the segment in which hotels

21 Both company-owned units (as leaders) and fran-chised units (as followers) may benefit from price lead-ership in concentrated markets. However, although fran-chised units in markets with only franchised units do notbenefit from price leadership, company-owned units inmarkets with only company-owned units do benefit.Therefore, we expect that franchised units on the averagebenefit less from concentration increases than company-owned units.

22 The same regression calculated within a five-mileradius rather than a two-mile radius yielded very similarresults except for the coefficient on the interaction be-tween royalty fee and concentration, which was positivebut insignificant.


were competing. It might be argued that hotels thatare in very different segments do not compete forthe same customers. We recalculated market-levelvariables, such as market concentration and meanoccupancy, assuming that hotels only competewith rivals in the same or adjacent segments. Theresults with this restricted market definition(model 11) are again largely consistent with previ-ous results (model 7). As before, company-ownedunits were observed to increase price as concen-tration increased, and franchised units did not.Royalty fees increased room rates, as expected,but this effect did not change depending on theconcentration.

In parallel with the quantitative analysis pre-sented above, we conducted interviews with indus-try participants, such as owner/managers of fran-chised hotels, managers of company-owned hotels,and top-level and midlevel managers of hotel cor-porations, to validate our assumptions, reasoning,and conclusions. One topic of these interviews wascompetition. Most interviewees agreed that compe-tition was a local phenomenon. Hotel managersseemed to know their competitors well. A salesmanager of a large company-owned hotel ex-plained, “I know my competitors personally. Wesend each other overflow customers.” We also dis-cussed pricing policies, focusing on the difference

between company-owned and franchised units.Pricing policies, practices, and the use of systemssuch as yield or revenue management varied signif-icantly among the different interviewees. Large,chain-affiliated hotels used yield management sys-tems more than small or independent hotels did.The interviews indicated that corporate headquar-ters took local demand and competitive conditionsinto account when putting pricing policies intoplace. Within the same hotel corporation, it seemedthat the local management of upscale brands hadmore discretion in pricing than the managers ofeconomy brands. Generally, however, local man-agement had limited discretion in price setting.One manager at corporate headquarters clarified:“We are very prudent with giving local manage-ment too much pricing discretion. In fact, localmanagement is closely monitored and gets re-warded if it correctly follows the revenue manage-ment system’s recommendations.” In contrast, fran-chisees had significantly more discretion in settingprices. In the words of a manager at the franchisingdepartment of a hotel corporation, “The chain willprovide guidelines for the price the franchisee cancharge to customers (a range). This ‘target’ pricerange depends on things like the brand, location,etc. However, if business is slow, the hotel ownercan decide to provide discounts. The franchisee is

TABLE 4Robustness Checks with Varying Definitions of Marketa

Independent VariablesModel 7:Zip Code

Model 10:Two-Mile

Radius

Model 11:Same orAdjacentSegment

Franchised ownership 0.00 (.00) 0.00 (.00) 0.00 (.01)Market concentration 0.12*** (.03) 0.21*** (.05) 0.05* (.02)Franchised ownership � market concentration �0.12*** (.03) �0.24*** (.05) �0.07** (.02)Franchised ownership � royalty fee 0.03** (.01) 0.02* (.01) 0.03** (.01)Franchised ownership � royalty fee � market concentration 0.09* (.04) 0.16* (.07) �0.05 (.03)Mean market occupancy 0.22*** (.01) 0.29*** (.01) 0.15*** (.01)Market capacityb �0.02** (.01) �0.02* (.01) �0.02** (.01)

Period fixed effects Yes Yes YesBrand fixed effects Yes Yes YesEstablishment fixed effects Yes Yes Yes

Observations 11,808 11,613 9,953Establishments 619 617 541R2 .97 .97 .97R2 (within) .26 .28 .26F 102*** 107*** 90***

a Monopoly hotels were excluded from the sample for the three models shown. Price (logarithm) is the dependent variable in all threemodels. Standard errors are in parentheses.

b Logarithm.* p � .05

** p � .01*** p � .001


free to do that.” These interviews confirmed ourargument that the hands of the company-ownedunit’s manager are tied—creating the possibility ofstrategic delegation—while franchisees have fulldiscretion and, therefore, commitment to priceleadership is unfeasible.

A final topic of our interviews was price leader-ship and collusion. There was widespread agree-ment that price increases could lead to higher prof-its. Although many interviewees agreed thatcollectively raising prices would increase profits,most did not see this happening in reality. Aninterviewed sales manager argued, “Collusion andprice setting doesn’t work. Someone will alwaysrenege. There is no commitment.” Another inter-viewee, vice president of a hotel corporation, sawcompany-owned units as playing a price leadershiprole in local markets: “After entry, we set a rela-tively high price and show that the market can bearthe higher price. Often, but not always, our com-petitors pick this up and increase their prices, too.”This last statement follows a reasoning that is iden-tical to this paper’s main argument: franchisedunits cannot credibly commit to raising prices inconcentrated markets, but company-owned unitscan increase the average market price through stra-tegic delegation and thus decrease the intensity ofrivalry.

It is worthwhile contrasting our findings withthose of previous research. Lafontaine and Slade(1997) cited six prior studies showing that fran-chised outlets charged a higher price than company-owned outlets, which was typically interpreted asevidence of double marginalization. Two studiesdealt with gasoline stations (Barron & Umbeck,1984; Shepard, 1993), two with fast-food chains(Lafontaine, 1995; Graddy, 1997), one with pubs(Slade, 1998a), and one with soft drink bottlers(Muris et al., 1992). Our results do not refute thesefindings. Specifically, model 1, controlling for mar-ket occupancy, market capacity, and period effects,shows that franchised units on average charge ahigher price than company-owned units. However,the cited studies—and model 1—do not distin-guish among several causal mechanisms that mightunderlie this price difference. The price differencebetween company-owned and franchised unitscould be a result of double marginalization, unob-served heterogeneity at the establishment levelcaused by the location or the quality of the productoffering, or, as we argue in the Theory section, thedifferent ability to credibly commit to price in-creases. Our empirical design allowed us to sepa-rate these potential explanations. First, we foundconfirming evidence for double marginalizationthrough the positive relationship between royalty

rates and prices of franchised units, making thisstudy the first to quantify the double marginaliza-tion effect and to tie it directly to royalty fees andcompetitive circumstances. Second, after control-ling for double marginalization, we found that inmore concentrated markets, company-owned unitsand their direct rivals priced higher than fran-chised units and achieved a higher performance,providing evidence for commitment via strategicdelegation. Finally, our empirical design allowedus to control for the third potential causal mecha-nism: unobserved heterogeneity at the establish-ment level.

Theoretical treatment of strategic delegationtheory has been ample; empirical evidence sup-porting this theory has been scant. Two notableexceptions are Slade (1998b) and Aggarwal andSamwick (1999). Slade found that the likeli-hood of vertical separation in gasoline stationsincreased as the benefit of separation throughstrategic delegation increased. Aggarwal andSamwick found, in a cross-sectional study ofcompanies listed on a stock exchange, that theweight placed on rivals in executive compensa-tion schemes decreased as concentration in-creased. The above findings provide preliminaryevidence for strategic delegation theory. In thisstudy, we show that chains may use company-owned units to credibly commit to less aggressivepricing behavior and soften competition in con-centrated markets, confirming strategic delega-tion theory. One way to understand our results isto look at them in light of a metaphor provided bythe well-known legend of Ulysses. When, in thecourse of their years at sea, Ulysses and his sail-ors approached the Sirens, whose beautiful sing-ing lured ships to run aground on jagged rocks, heordered them to tie him to the mast of the shipand to put wax in their own ears to avoid beingtempted by the Sirens’ songs. By doing so, Ul-ysses managed to steer away from the Sirens andcontinue his journey. Equivalently, by tying thehands of managers, one may be able to make themresist the short-term temptations of price compe-tition in their market and achieve a more profit-able outcome.23

The direct generalizability of our findings is lim-ited to industries in which both franchised andcompany-owned establishments compete in a lo-cally oligopolistic context, which nevertheless in-cludes most retail and service industries (Greve &

23 Elster (1979) discussed this example of the idea thatindividuals can sometimes choose their own constraintsin his Ulysses and the Sirens.


Baum, 2001). Moreover, the effect whereby strate-gic delegation makes company-owned units behaveless aggressively than franchised units is contin-gent on an industry being based on price competi-tion among differentiated offerings. The hypothe-ses may not extend to other industry settings inwhich competition focuses on capacity or invest-ment preemption and where a more aggressive pre-emptive strategy is appropriate. However, the gen-eral finding that the ownership and organizationalcharacteristics of a firm affect competition betweenfirms may be more broadly applicable. This studytook firm ownership as one characteristic that af-fects managerial incentives, focusing on the fran-chise versus company-owned distinction. Futureresearch could expand this exploration to otherownership typologies, such as government-ownedfirms, employee cooperatives, family-owned firms,management buyouts, and so forth. Moreover, be-fore one could conclude that in general “the insidematters for competition,” future research shouldstudy other organizational characteristics, such ascompensation schemes and organizational struc-ture, and their effect on interfirm rivalry.

In many industries, companies make decisionsabout organizational forms and boundaries: for ex-ample, whether to own or to use third-party chan-nels, whether to enter markets through joint ven-tures or acquisitions, or whether to consolidatethrough alliances or mergers. This study allowscompanies facing these choices to understand howdifferent organizational arrangements may affecttheir ability to capture value in the implicit nego-tiation with rivals and customers. We find thatcompanies that have full control over local unitsand can implement administrative systems thatconstrain managerial behavior (company owner-ship) are better able to commit to price leadership.This finding is not only theoretically important butalso practically relevant, because in many cases acompany’s ability to manage its competitive envi-ronment is an important determinant of itsperformance.

REFERENCES

Aggarwal, R. K., & Samwick, A. A. 1999. Executive com-pensation, strategic competition, and relative perfor-mance evaluation: Theory and evidence. Journal ofFinance, 54: 1999–2043.

Aiken, L. S., & West, S. G. 1991. Multiple regression:Testing and interpreting interactions. NewburyPark, CA: Sage.

Barron, J. M., & Umbeck, J. R. 1984. The effects of differ-ent contractual arrangements: The case of retail gas-

oline markets. Journal of Law & Economics, 27:313–328.

Baum, J. A. C., & Mezias, S. J. 1992. Localized competi-tion and organizational failure in the Manhattan ho-tel industry, 1898–1990. Administrative ScienceQuarterly, 37: 580–604.

Bergen, M., Dutta, S., & Walker, O. C. 1992. Agencyrelationships in marketing: A review of the implica-tions and applications of agency and related theo-ries. Journal of Marketing, 56(3): 1–24.

Besanko, D., Dranove, D., & Shanley, M. 2000. Economicsof strategy (2nd ed.). New York: Wiley.

Blair, R. D., Lafontaine, F. 1999. Will Khan foster orhinder franchising? An economic analysis of maxi-mum resale price maintenance. Journal of PublicPolicy & Marketing, 18(1): 25–36.

Bonanno, G., & Vickers, J. 1988. Vertical separation. Jour-nal of Industrial Economics, 36: 257–265.

Bradach, J. L. 1998. Franchise organizations. Boston:Harvard Business School Press.

Brickley, J. A., & Dark, F. H. 1987. The choice of organi-zational form: The case of franchising. Journal ofFinancial Economics, 18: 401–420.

Bulow, J. I., Geanakoplos, J. D., & Klemperer, P. D. 1985.Multimarket oligopoly: Strategic substitutes andcomplements. Journal of Political Economy, 93:488–512.

Canina, L., Enz, C. A., & Harrison, J. S. 2005. Agglomer-ation effects and strategic orientations: Evidencefrom the U.S. lodging industry. Academy of Man-agement Journal, 48: 565–581.

Caves, R. E., & Murphy, W. F. 1976. Franchising: Firms,markets, and intangible assets. Southern EconomicJournal, 42: 572–586.

Chandy, R. K., & Tellis, G. J. 1998. Organizing for radicalproduct innovation: The overlooked role of willing-ness to cannibalize. Journal of Marketing Re-search, 35: 474–487.

Christensen, C. M. 1997. The innovator’s dilemma:When new technologies cause great firms to fail.Boston: Harvard Business School Press.

Chung, W., & Kalnins, A. 2001. Agglomeration effectsand performance: A test of the Texas lodging indus-try. Strategic Management Journal, 22: 969–988.

Combs, J. G., Michael, S. C., & Castrogiovanni, G. J. 2004.Franchising: A review and avenues to greater theo-retical diversity. Journal of Management, 30: 907–932.

Conlin, M. 2002. The effect of franchising on competi-tion: An empirical analysis. Working paper, Syra-cuse University.

Conlin, M., & Kadiyali, V. 2006. Entry deterrence ca-pacity in the Texas lodging industry. Journal ofEconomics and Management Strategy, 15: 167–185.


Directory of Hotel & Lodging Companies. 2003.Washington, DC: American Hotel & Lodging Asso-ciation.

Dixit, A. 1980. The role of investment in entry-deter-rence. Economic Journal, 90(March): 95–106.

Dixit, A., & Nalebuff, B. J. 1991. Thinking strategically:The competitive edge in business, politics, and ev-eryday life. New York: Norton.

Elster, J. 1979. Ulysses and the sirens. Cambridge, U.K.:Cambridge University Press.

Ferrier, W., Smith, K., & Grimm, C. 1999. The role ofcompetitive action in market share erosion and in-dustry dethronement: A study of industry leadersand challengers. Academy of Management Journal,42: 372–388.

Fershtman, C., & Judd, K. L. 1987. Equilibrium incentivesin oligopoly. American Economic Review, 77: 927–940.

Fudenberg, D., & Tirole, J. 1984. The fat-cat effect, thepuppy-dog ploy, and the lean and hungry look.American Economic Review, 74: 361–366.

Ghemawat, P. 1991. Commitment: The dynamic of strat-egy. New York: Free Press.

Graddy, K. 1997. Do fast-food chains price discriminateon the race? Journal of Business & Economic Sta-tistics, 15: 391–401.

Greve, H. R. 2003. Why are there so many multi-unit organizations? Strategic Organization, 1:109 –115.

Greve, H. R., & Baum, J. A. C. 2001. Introduction: Amultiunit, multimarket world. In H. R. Greve &J. A. C. Baum (Eds.), Advances in strategic man-agement, vol. 18—Multiunit organization andmultimarket strategy: 1–28. Greenwich, CT.

Greve, H. R., & Goldeng, E. 2004. Longitudinal analysis instrategic management. In D. D. Bergh & D. Ketchen(Eds.), Research methods in strategy and manage-ment: 135–163. Amsterdam: Elsevier.

Gupta, A. K., & Govindarajan, V. 1984. Business unitstrategy, managerial characteristics, and businessunit effectiveness at strategy implementation. Acad-emy of Management Journal, 27: 25–41.

Hamilton, B. H., & Nickerson, J. A. 2003. Correcting forendogeneity in strategic management research. Stra-tegic Organization, 1: 51–72.

Hansmann, H., Kessler, D., & McClellan, M. 2003. Own-ership form and trapped capital in the hospital in-dustry. In E. L. Glaeser (Ed.), The governance ofnot-for-profit organizations: 45–70. Chicago: Uni-versity of Chicago Press.

Hart, O. 1995. Firms, contracts, and financial struc-tures. Oxford, U.K.: Oxford University Press.

Ingram, P. 1996. Organizational form as a solution tothe problem of credible commitment: The evolu-tion of naming strategies among U.S. hotel chains,

1896 –1980. Strategic Management Journal, 17:85–98.

Ingram, P., & Baum, J. A. C. 1997. Chain affiliation andthe failure of Manhattan hotels, 1898–1980. Admin-istrative Science Quarterly, 42: 68–102.

Kalnins, A. 2004. An empirical analysis of territorialencroachment within franchised and company-owned branded chains. Marketing Science, 23:476–489.

Kehoe, M. R. 1996. Franchising, agency problems, andthe cost of capital. Applied Economics, 28: 1485–1493.

Lafontaine, F. 1992. Agency theory and franchising:Some empirical results. Rand Journal of Econom-ics, 23: 263–284.

Lafontaine, F. 1995. Pricing decisions in franchisedchains: A look at the fast-food industry (Workingpaper no. 5247). Cambridge, MA: National Bureau ofEconomic Research.

Lafontaine, F. 1999. Franchising versus corporate own-ership: The effect on price dispersion. Journal ofBusiness Venturing, 14: 17–34.

Lafontaine, F., & Shaw, K. L. 1999. The dynamics offranchise contracting: Evidence from panel data.Journal of Political Economy, 107: 1041–1080.

Lafontaine, F., & Slade, M. E. 1997. Retail contracting:Theory and practice. Journal of Industrial Econom-ics, 45(1): 1–25.

Muris, T. J., Scheffman, D. T., & Spiller, P. T. 1992.Strategy and transaction costs: The organization ofdistribution in the carbonated soft drink industry.Journal of Economics and Management Strategy,1: 83–128.

Perry, M. K. 1989. Vertical integration: Determinants andeffects. In R. Schmalensee & R. D. Willig (Eds.),Handbook of industrial organization, vol. I: 183–255. Amsterdam: Elsevier Science Publishers.

Schelling, T. C. 1960. The strategy of conflict. Cam-bridge, MA: Harvard University Press.

Schmidt, T. 1994. An analysis of intrabrand competitionin the franchise industry. Review of Industrial Or-ganization, 9: 293–310.

Shepard, A. 1993. Contractual form, retail price, andasset characteristics in gasoline retailing. Rand Jour-nal of Economics, 24: 58–77.

Sklivas, S. D. 1987. The strategic choice of managerialincentives. Rand Journal of Economics, 18: 452–458.

Slade, M. E. 1998a. Beer and the tie: Did divestiture ofbrewer-owned public houses lead to higher beerprices? Economic Journal, 108: 565–602.

Slade, M. E. 1998b. Strategic motives for vertical sep-aration: Evidence from retail gasoline markets.Journal of Law, Economics & Organization, 14:84 –113.


Spengler, L. 1950. Vertical integration and anti-trust pol-icy. Journal of Political Economy, 58: 347–352.

Stigler, G. J. 1964. A theory of oligopoly. Journal ofPolitical Economy, 72: 44–61.

Tirole, J. 1988. The theory of industrial organization.Cambridge, MA: MIT Press.

Vickers, J. 1985. Delegation and the theory of the firm.Economic Journal, 95(supplement): 138–147.

Williamson, O. E. 1991. Strategizing, economizing, andeconomic organization. Strategic ManagementJournal, 12(winter special issue): 75–94.

Yeap, C. A. 2006. Residual claims and incentives inrestaurant chains. Working paper, University ofMinnesota, Minneapolis.

Yin, X., & Zajac, E. J. 2004. The strategy/governancestructure fit relationship: Theory and evidence infranchising arrangements. Strategic ManagementJournal, 25: 365–383.

Govert Vroom ([email protected]) is an assistant pro-fessor of management at the Krannert School of Manage-ment, Purdue University. He earned his Ph.D. in man-agement from INSEAD. His research interests includecompetitive strategy, managerial incentives, organization-al design, and industrial organization.

Javier Gimeno ([email protected]) is a professorof strategy at INSEAD. He earned his Ph.D. in strategicmanagement from Purdue University. His research inter-ests include competitive strategy, competitive dynamicsin multimarket contexts, the effect of organizational in-centives and delegation on competitive interaction, andentrepreneurship.


ownership form, managerial incentives, and the intensity of rivalry

Documents