1 Introduction

Franchising is a type of inter-firm collaboration that is widely used in global business activity. In it, a firm, the franchisor, grants the right to use its business concept to legally independent firms, the franchisees. In the US, it accounts for nearly $1.3 trillion of sales from 453,326 businesses in 295 industries and employs nearly 7.9 million workers (US Census Bureau 2007). Although the figures are smaller in Spain, franchising is growing constantly, reaching 15.2 % of Spanish retail sales and employing 400,000 people (Tormo & Asociados 2009).

However, franchise relationships present a conflict of interest between the parties (Rubin 1978; Brickley and Dark 1987; Lafontaine 1992). Given that this cooperation requires a joint effort and costly mutual monitoring, it is feasible that parties may engage in opportunistic behavior, seeking to maximize personal gain at the expense of the partner.

The effectiveness of franchising as a successful organizational form will depend on its capacity for limiting such opportunistic behavior. Transaction Cost Economics (Klein et al. 1978; Macneil 1978; Williamson 1985; Klein 1995) suggests that parties design their contracts, introducing different safeguards, to attenuate it. The franchise contract literature has made great efforts to explain the economic rationality of these contracts (Brickley and Dark 1987; Brickley et al. 1991) and has outlined the roles and responsibilities of each party, the allocation of decision and control rights, the planning for various contingencies, how the parties will communicate and how to resolve disputes (Argyres and Mayer 2007). Many empirical papers support the idea that the contingencies introduced in the franchise contract mitigate the contractual hazards present in such relationships (Lafontaine 1992; Arruñada et al. 2001). Similarly, Lafontaine and Slade (2013) review empirical papers looking for the effects of different contractual arrangements on firm performance. They show evidence of several significant effects.

However, it is not so clear in the literature which mechanisms of governance should be used, and when, to be successful either in franchising or in any other inter-firm relationship. Several authors have concluded that choosing the right mix of formal and relational governance mechanisms affects performance (Luo 2002; Poppo and Zenger 2002; Gulati and Nickerson 2008; Hoetker and Mellewigt 2009). Both mechanisms afford benefits: formal contracts imply the ex ante specification of obligations, rewards, risks, procedures and so forth through individual contractual provisions. This allows for many potential conflicts to be solved in advance by mutual agreement. Relational contracting hardly incurs any ex ante writing costs because it is largely based on trust and social identification (Dyer and Singh 1998), so flexibility is enhanced and ex post opportunism (renegotiation costs) is reduced because parties are interested in preserving long-term or future profitable transactions (Mesquita and Brush 2008).

Another issue in the literature is to understand the allocation of property/decision rights and when they have to be specified in greater detail in contracts. At least three theoretical arguments explain the asymmetrical allocation of decision rights and the different degree of detail in clause specifications. First, literature on organizational learning (Lieberman 1984; Argote 1999; Mayer and Argyres 2004; Ryall and Sampson 2006) suggests that firms slowly learn to contract on the basis of their experience and past problems. This experience in understanding the potential for conflicts in their relations and knowledge about how to efficiently resolve parties’ disputes through the contract can be a valuable capacity affecting firms’ performance. Differences in these contract design capabilities explain why some experienced companies introduce more details and contingencies in their contracts (Argyres et al. 2007; Cochet and Garg 2008): companies without these capabilities are not able to design such detailed contracts, so are more exposed to conflicts which could reduce their profitability.

Second, franchisors usually jeopardize much more reputational capital in the relationship than franchisees, making them more sensitive to franchisees’ opportunism (Arruñada et al. 2001). Given that it has been argued that self-enforcement mechanisms such as reputational capital substitute contract formalization (Macaulay 1963; Gulati 1995; Zaheer and Venkatraman 1995; Dyer and Singh 1998), we expect to see franchisors interested in detailed contracts since formalization turns into their only tool to attenuate franchisees’ opportunism. Conversely, franchisors’ reputation may also assure franchisees of their fair behavior (Klein 1996; Williamson 1983), making it unnecessary to formalize this part of the contract.

Third, contract formalization is a negotiation process between parties in which bargaining power matters.Footnote 1 Assuming that the franchisor is the best-established party (Klein 1980), greater detail in franchisee obligations may be simply because franchisors force franchisees to accept their conditions in take-it-or-leave-it negotiations. Similarly, less successful franchisors, who have less knowledge about their business and a brand that is not well-known in the market, are unlikely to have so much power for imposing their conditions on potential franchisees, and this will affect their contract structure. Michael (2000a) argues that using plural forms, selecting inexperienced franchisees, and adopting a long training program are devices for increasing the franchisor’s bargaining power. Conversely, Argyres and Bercovitz (forthcoming) show that the presence of independent franchisee associations influences contract design because they increase franchisees’ bargaining power.

Given all this previous research showing that contract design affects firm performance, the aim of this paper is to identify differences in contract design between successful and less successful franchise chains. More precisely, we are particularly interested in comparing a) the degree of contractual completeness, defined as the extent to which contingencies are specified in contracts (Luo 2002; Mesquita and Brush 2008; Vanneste and Puranam 2010)Footnote 2 and b) the specific content of contracts from both the franchisor and franchisee perspectives, identifying parties’ obligations and rights.

We help fill a gap in the literature. While the relevance of completeness has been considered in other types of contract such as joint ventures (Luo 2002), strategic alliances (Reuer et al. 2006) and outsourcing (Mesquita and Brush 2008), the literature on franchising has not paid much attention to this point. Additionally, we are not aware of any study showing the design differences between contracts used by successful and less successful franchise chains. There are works focused on determining how certain contractual clauses affect performance (Agrawal and Lal 1995; Shane 1998; Michael 2000b; Azoulay and Shane 2001; Shane et al. 2006; Michael and Combs 2008; Lafontaine and Slade 2013), but they do not consider the contract as a whole or the different degree of detail in the contract specifications (i.e. relating contractual completeness with chain success, or assessing which type of franchisee obligations are most relevant).

The remainder of the article is structured as follows. After this introduction, the second section describes the data collection process, information sources and econometric models, then reports and discusses the results. Finally, some conclusions are given.

2 Empirical evidence

2.1 Data and cluster analysis

Given that we are interested in identifying the differences in terms of contract design between successful and less successful franchise chains, we contacted all Spanish franchise chains and asked them to collaborate in our study.Footnote 3 We decided not to include foreign franchisors operating in Spain in order to avoid a potential bias caused by the effect of their national regulations on contract design. It is common practice for franchise chains in foreign markets to use the same contract as in their domestic markets (Lafontaine and Oxley 2004) so, if the sample included both Spanish and foreign franchisors, the contractual differences might be influenced by the law of the franchisor’s country of origin.Footnote 4

We contacted 805 franchisors and requested them to send information about their companies and particularly their franchise contracts. 293 franchisors agreed to collaborate. Despite our efforts in following several standard recommendations in the literature to increase the response rate,Footnote 5 many of them did not send the information requested on their contracts. We finally closed the request for information having received 74 contracts, from both the services and retail industries (9.2 % overall response rate). The services included real estate agencies, hairdressers, travel agencies or consultancies, and retailing included catering and clothing establishments or vending operations. The information needed to compare contractual design in franchise chains was drawn directly from these 74 contracts between franchisors and franchisees.

We complemented the contract information with general information about the chains and their success. We used different sources of data: the dossier package sent by them, the franchisors’ web sites or, when neither of these were available, the Professional Franchise Guides. Financial information about franchisors was obtained from the SABI data base (Bureau van Dijk), which gives the net income statement and balance sheet for all companies operating in Spain.

To test for a potential response bias in our sample, we followed the Armstrong and Overton (1977) procedure. We compared several variables in early-returned questionnaires and late-returned questionnaires. This comparison assumes that late respondents share similar characteristics and response biases with non-respondents. Analyses indicated that no significant mean differences existed between early and late respondents regarding completeness. Furthermore, we compared the industries represented in the sample with the population (Poppo and Zenger 2002). 24 different industries are represented in our sample, with hospitality and fashion being the largest. These data are the same as those for the Spanish franchise population, as given in the reports drawn up by the Spanish Franchise Association and Tormo & Asociados. Therefore, the sample and population did not appear to differ by industries.

We then classified the franchise chains according to their success in the market. We used three different variables for success, YEARS, SIZE and ROA. YEARS indicates the number of years that a chain has been working as a franchise, and is an indicator of its survival in the market (Bates 1998; Lafontaine and Shaw 1998; Shane and Foo 1999). SIZE is the total number of establishments held by each chain and proxies the relevance of the franchisor’s brand name in the market (Lafontaine 1992; Agrawal and Lal 1995; Arruñada et al. 2001) and, therefore, its success. Return on assets (ROA), defined as Operating results/Total assets, indicates the company’s performance achieved as a result of its investment in assets, irrespective of its financial structure. This is an indicator of the chain’s profitability in the short run (Combs and Ketchen 1999; Combs et al. 2004; Arruñada et al. 2009).

Using these three measures of success, we performed a two-step cluster analysisFootnote 6 which identified the existence of two groups of franchise chains (Table 1): “less successful chains” (group 1) and “successful chains” (group 2). The chains forming part of the latter group were those that had been franchising their businesses longest, had a higher number of establishments and achieved the best financial performance, as shown in Table 1.

Table 1 Group profiles: chains’ success
Full size table

2.2 Results

Comparing these two groups of companies in terms of contractual design requires, first, measuring the degree of completeness of their contracts as a whole and, second, analyzing their specific content from both the franchisor and franchisee perspectives, identifying parties’ obligations and rights.

We used four variables for contractual completeness: the number of CONTINGENCIES included in each contract (each contingency represents a specific aspect or potential problem in the franchise relationship),Footnote 7 the number of PAGES, the number of WORDS and the degree of DETAIL of each contract relative to the rest of the contracts in the sample.Footnote 8 The average contract has 60 contingencies, 18 pages, over 6000 words and a score of 18.15 for detail. The CONTINGENCIES variable especially improves previous measures of completeness because we directly measure the number of contingencies and consider all of them when estimating the level of completeness. Authors such as Poppo and Zenger (2002) and Hendrikse and Windsperger (2011) measure this concept indirectly because they ask managers to indicate on a Likert scale to what extent the formal contract was highly customized or detailed. Other works only consider certain clauses or contingencies for estimating this measure (Parkhe 1993; Luo 2002; Reuer et al. 2006; Reuer and Ariño 2007; Mesquita and Brush 2008).

Table 2 shows that the successful chains are those with the most complete contracts, covering a larger number of contingencies, with more pages and words and greater detail. However, the differences with regard to less successful chains are only statistically significant for the number of contingencies. While less successful chains include an average of 57.71 contingencies in their contracts, the most successful ones include 65.42 contingencies. This suggests that, on average, successful chains identify and negotiate ex ante more potential problems than less successful companies. This could be because they have more experience (see Table 1: 14.67 vs. 5.16 years franchising) which has helped them learn how to manage and solve any conflicts in the relationship. They progressively improve their knowledge about relevant aspects and details which they then formalize in their contracts (contingencies). This knowledge about what to include in the contract, and how, is what the literature calls contract design capabilities (Mayer and Argyres 2004; Argyres et al. 2007; Ryall and Sampson 2009; Vanneste and Puranam 2010). The literature on organizational learning (Lieberman 1984; Argote 1999; Mayer and Argyres 2004) and, to a lesser extent, transaction cost theory (Williamson 1985) maintain that this higher formalization improves performance because the inclusion of more contingencies in the contract places more limits on the action of the parties, which facilitates ex post enforcement. This reduces renegotiation costs and helps to mitigate potential opportunism (Klein et al. 1978; Williamson 1985; Poppo and Zenger 2002).

Table 2 Contractual completeness
Full size table

We then went a step further, analyzing what kinds of contingencies are included when contracts become more complete. Table 3 shows part of this analysis, distinguishing five contingency categories or groups: franchisor’s obligations (e.g. to provide the method, to assist the franchisees, to provide training…), franchisees’ obligations (e.g. to respect geographical restrictions, to adapt the establishment, compliance with the method, …), grounds for termination in favor of the franchisor (e.g. franchisee’s failure to pay, non-fulfillment of quality standards, use of unauthorized suppliers…), grounds for termination in favor of the franchisee (e.g. the franchisor does not provide the committed training or assistance, the franchisor declares insolvency or bankruptcy, the franchisor passes away, …) and other contingencies (which cannot be classified as either rights or obligations of either the franchisor or the franchisee, such as contract duration).Footnote 9 From this table, we can extract two main findings.

The first is that, irrespective of chain success, contracts cover franchisors’ rights more than franchisees’ rights. Table 3 shows that, on average, 19.42 franchisee obligations are included in contracts against 8.83 franchisor obligations. This difference is statistically significant. Additionally, more detailed grounds for termination are included in favor of the franchisor than in favor of the franchisees (16.42 vs. 5.04), and this difference is also statistically significant. In other words, franchisees’ obligations and grounds for termination in favor of the franchisor, and therefore against the franchisee, are specified in contracts more. This different treatment of parties’ obligations/rights has been already noted in the literature (Klein 1980; Al-Najjar 1995; Arruñada et al. 2001; Spencer 2008). Klein (1980, p. 360) suggested an explanation. He states that “when both parties can cheat, explicit contractual restraints are often placed on the smaller, less well-established party (the franchisee), while an implicit brand name contract-enforcement (self-enforcing) mechanism is relied on to prevent cheating by the larger, more well-established party (the franchisor)”. However, another explanation for this result could be the existence of unequal bargaining power between the franchisor and the franchisee (Schwartz 1974; Klein 1980; Lagarias and Boulter 2010). The franchisor has more bargaining power than franchisees because he is usually the better-established party in the relationship: it is the franchisor’s business that is being traded and it is the franchisor who draws up the contract and writes the operations manual. The greater detail regarding franchisee obligations may be simply because franchisors force franchisees to accept their conditions in take-it-or-leave it negotiations. Unfortunately, our analysis does not enable us to disentangle which argument is the most relevant.

The second finding is that successful franchise chains seem to introduce more contingencies on average than less successful ones, either in terms of parties’ obligations or in terms of termination rights. While less successful chains include an average of 8.42 contingencies regarding the franchisor’s obligations and 18.21 regarding the franchisee’s obligations, the most successful chains include 9.30 and 20.82 contingencies, respectively. A similar difference is observed regarding the grounds for termination. Successful chains introduce more termination rights in favor of both the franchisor (17.30 grounds) and the franchisee (5.27 grounds) than less successful chains (15.66 and 4.84 respectively). However, franchisees’ obligations is the only category in which we observe statistically significant differences. This means that the observed difference in terms of completeness is actually because successful chains specify franchisees’ obligations in more detail than less successful chains. The other contingency categories (franchisor’s obligations, grounds for termination and other contingencies) are not statistically different.

Although we cannot assume causality in our empirical test, these results are consistent with previous empirical studies and suggest that franchise chains have to pay more attention to franchisees’ than to franchisors’ behavior to be successful. On the one hand, the fact that less successful chains introduce such contingencies in their contracts less may be due to the fact that they refer to problems that are not easy to identify a priori. In order to include them in the contract it is necessary to know your business well, that is, to have developed learning about your business.Footnote 10 Indeed, the literature on learning states that firms learn about potential conflicts and hazards slowly and incrementally, introducing them in contracts as they experience these problems (Cyert and March 1963). That is, rather than anticipating such conflicts, the parties have to actually experience an adverse situation before addressing it in a new contract because attempts to foresee contracting hazards and incentive problems in contracts are inadequate, requiring them to be added in subsequent contracts (Mayer and Argyres 2004).

On the other hand, Arruñada et al. (2001) suggests for car dealership contracts that the higher a manufacturer’s reputational capital (a manufacturer plays a similar role to a franchisor), the higher the “contractual asymmetry” between manufacturers and car dealers (i.e. between franchisees and franchisors in our contracts). They argue that the damage that car dealers’ (franchisees) opportunistic behavior can inflict on manufacturers (franchisors) is greater when the latter have higher reputational capital. If we assume that successful chains have more reputational capital, our finding is consistent with this argument. This reputational capital argument cannot be reversed towards franchisees. Unlike franchisors, franchisees are usually small entrepreneurs whose reputational capital is very limited, so it does not serve as a guarantee of good behavior for the franchisor. Consequently, the only mechanism available for the franchisor to try to exert greater control over franchisees is the contract, including more contingencies to ensure that its instructions are followed (Mellewigt et al. 2007).

Finally, this difference between successful and less successful franchise systems could also be explained in terms of bargaining power. The idea is that less successful franchisors probably have less bargaining power than successful franchisors. As a consequence, less successful franchisors are not able to impose their conditions on their franchisees while successful franchisors can. This situation enhances the above argument regarding asymmetry in contractual obligations between franchisee and franchisor.

Table 3 Franchisors’ and franchisees’ obligations, grounds for termination and other contingencies
Full size table

What franchisees’ obligations are included in the contracts? Table 4 compares the absolute and relative frequencies of occurrence (i.e. inclusion in the contract) between successful and less successful franchise chains of 40 contingencies regarding franchisees’ obligations. We have differentiated three contingency categories: business-concept contingencies, i.e. contingencies related to outlet operations management and corporate image; financial conditions, i.e. contingencies about investments and payments between parties; and post-contractual contingencies, i.e. contingencies about parties’ behavior after contract termination. Contingencies are ranked according to the relative frequency of occurrence by less successful chains within each category. First, we observe that 31 contingencies (77.5 % of the total) are introduced by successful chains more frequently than by less successful ones. This backs the idea expressed above that successful chains make their contracts more complete.

Table 4 Franchisees’ obligations: comparison of groups
Full size table

Second, statistically significant differences between successful and less successful chains occur more frequently in business-concept contingencies than in financial and post-contractual contingencies. More than one out of three (37.04 %) business-concept contingencies are statistically significant, while none of the financial contingencies show significant differences between successful and less successful franchise chains and only one out of four post-contractual contingencies. This suggests that contract differentiation comes from the introduction of contingencies related to day-to-day business operations more than from those related to financial aspects or problems after contract termination. Such business operation problems are usually related to the two big issues that the literature on franchise chains has largely identified: the provision of efficient incentives to the person responsible for the outlet (Lafontaine 1992), and keeping the uniformity of the business throughout the chain to reduce consumers’ search costs (Bradach 1998).

What are the specific contingencies that are not used sufficiently by less successful chains? Table 4 shows that, within the business-concept contingencies, there is one set of contingencies that are highly used (relative frequency over 61 %) by all chains, but in which differences between successful and less successful chains are statistically significant. They are the franchisee’s obligation to adapt the establishment, to comply with the method, to exclusively offer/use authorized products/suppliers, to hand over documentation and financial statements, not to launch unauthorized advertising campaigns, to underwrite liability and other insurance policies. There is a second set of contingencies that are less used (relative frequency under 21 %) but that still show statically significant differences. They are the franchisees’ obligation to notify the franchisor of any improvement in know-how, to buy a minimum amount from the central office and to run the business personally. These latter contingencies are also the ones with the highest average growth from the less successful franchise group to the successful group.

Taking successful chain contracts as a benchmarking reference, these results suggest that less successful chains have to introduce more contingencies related with the business-concept in their contracts. In particular, it seems efficient to restrict franchisees’ decision rights. All such business-concept contingencies aim to guarantee that the franchisee replicates the franchised business operations exactly in its outlet, and it seems that successful chains achieve this by compelling the franchisee to comply with the method, to use the same inputs/outputs and to adopt exactly the same corporate image as the franchisor.

Table 4 also indicates that there are two post-contractual franchisee obligations which show statically significant differences between successful and less successful chains. They are related to the prohibition of continuing to use the franchisor’s method and to the franchisor’s preference for the franchisee’s stocks. The financial conditions are not statistically different between the groups.

Finally, we do not observe differences between successful and less successful chains in terms of the proportion of chains that include specific franchisor’s obligations (Table 5). We argue that the low level of formalization (i.e. contingencies) of franchisor’s obligations may be either because it has greater bargaining power or because its reputation and brand name act as a guarantee for franchisees of its fair behavior (Klein 1996; Williamson 1983). Again, assuming that success is related to franchisors’ reputational capital, our results do not support the idea that self-enforcement mechanisms such as reputation substitute contract formalization, as several authors have emphasized (Macaulay 1963; Gulati 1995; Zaheer and Venkatraman 1995; Dyer and Singh 1998).

Table 5 Franchisor’s obligations: comparison of groups
Full size table

3 Conclusions

We compare successful and less successful franchise chains in terms of contract design, particularly related to the degree of completeness and to the specific content of contracts from both the franchisor and franchisee perspectives. We contribute to the contract literature by filling a gap in the attention paid to completeness in franchise contracts and by exploring which’ contingencies are relevant for inclusion in contracts.

Several theoretical arguments justify asymmetric contractual design between franchisor and franchisee and between successful and less successful chains. On the one hand, several transaction cost economics arguments explain this asymmetry in terms of providing efficient incentives to both parties. Similarly, the organizational learning literature explains these differences in reference to different parties’ experiences and contractual capacities. Finally, greater bargaining power for franchisors may also explain why their obligations are not as detailed as franchisees’ ones. Unfortunately, our empirical analysis does not allow these effects to be untangled.

The empirical analysis was performed on a sample of 74 Spanish franchise contracts. Given that contractual completeness is a theoretical concept and is almost impossible to measure empirically, we defined contractual completeness as the extent to which contingencies are specified in contracts and we assume that the more contingencies a contract covers, the more complete it is. The success of each chain was decided by means of two-step cluster analysis using the chain size, the number of years franchising and the return on assets as measures of chain performance.

The results suggest that franchise contracts are unbalanced: contracts cover franchisees’ obligations more than franchisors’ obligations, irrespective of the chain’s success. When we compare successful and less successful franchise systems, we observe that successful chains are those with the most complete contracts, covering a larger number of contingencies. We also see that they contain more pages and words and greater detail but these differences are not statistically significant.

We then analyzed what kinds of contingencies are included when contracts become more complete. We highlight three findings. First, franchisee obligations is the only category of contingencies in which we observe statistically significant differences between successful and less successful franchise chains. This indicates that the observed difference in terms of completeness is because successful chains specify their franchisees’ obligations in more detail than less successful chains. The other contingency categories (franchisor’s obligations, grounds for termination and other contingencies) are not statistically different.

Second, the relative frequency of occurrence of franchisees’ obligations varies substantially. Statistically significant differences between successful and less successful chains occur more frequently in business-concept contingencies than in financial and post-contractual contingencies. This suggests that contract differentiation in successful franchise systems lies in the introduction of contingencies related to day-to-day business operations more than in those relating to financial aspects (which are not statistically significant) or problems after contract termination.

Third, we do not observe differences in the relative frequency of occurrence of franchisor obligations between successful and less successful chains. If we assume that successful franchise chains have greater reputational capital, this finding suggests that franchisors’ reputational capital (i.e. a self-enforcement mechanism) does not affect the formalization of their obligations, so does not support a substitution effect between formalization and reputational capital.

As for practitioners, our results suggest that chains have to pay much more attention to franchisees’ than to franchisors’ behavior. Furthermore, if successful chain contracts can be taken as a benchmark, these results suggest that franchise systems have to introduce in their contracts more contingencies related to franchisees’ obligations and, more specifically, to the business-concept if they are to improve. It seems efficient to restrict franchisees’ decision rights, especially in terms of purchasing and procurement, method, product range and image.

This study is not without limitations. As mentioned above, differences in franchise contract design between successful and less successful chains can be explained using different theoretical arguments. However our analysis does not enable us to disentangle to what extent and under what circumstances each of these arguments is relevant. This limitation suggests there is a need for quantitative research, which is an important item on our research agenda.