Relationships in Franchised Distribution System: the Case of the Spanish Market

Abstract

The purpose of this article is to examine long-term inter-firm relationship management and to investigate empirically if trust and commitment are essential variables for relationship success. This study analyses both the relationship between two behaviours (cooperation and exchange information) and the development of trust and commitment as well as their effects on long-term orientation, satisfaction and performance. The unit of analysis is the dyadic relationship, assessing separately franchisor's and franchisees' perspectives, and comparing similarities and differences across perspectives (Anderson & Narus, 1990; Jap, 1999). The model proposed has been tested with 107 franchisors and 102 franchisees operating both in the retail (food and beverages, fashions, furniture …) and the service sector (estate agencies, education, dry-cleaner's, diverse services …) in the Spanish franchised distribution system. Results show that both commitment and trust are key variables for long-term relationship success from the franchisors perspective, but that from the franchisees side, trust is the key variable for relationship success.

Keywords:

• Franchising
• trust
• commitment
• structural equations model
• distribution
• Spain

Acknowledgements

The authors wish to express their gratitude for the financial help received from the Government of Aragon through the GENERES project (Ref. S09 (26779)) and from the Science and Technology Ministry of Spain by means of the CICYT project (Ref. SEC 2002-03949).

Notes

They measured benevolence and honesty separately and employed its average as a global measure of trust.

Doney and Cannon (1997) verified that benevolence and credibility are highly correlated, although they are conceptually different. Thus, they could not discriminate these dimensions empirically.

In order to facilitate the reading of the hypotheses, it is sufficient to replace the term franchisor by franchisee and vice versa.

Anderson and Narus (1990) considered that cooperation yielded past experiences, while trust tapped present state, that they had to change their model considering cooperation as an antecedent of trust.

If we analyze separately affective and calculative commitment, trust has a positive effect on affective commitment and a negative influence on calculative commitment (Geykens and Steenkamp, 1995; Geykens et al., 1996).

Ganesan (1994) showed a significant and positive relation between credibility and long-term orientation, but he did not obtain a significant relation between benevolence and long-term orientation.

Some authors focus on the relationship between trust and economic satisfaction (Mohr & Spekman, 1994; Siguaw, Simpson & Baker, 1998).

The chi square contrast for franchisors was carried out considering that in Spain 57.03% of franchisors companies operate in retailing and the rest (37.96%) in the service sector. Results revealed a value of the statistic 0.832 (df = 1), associated with a p-value of 0.362. From the franchisees perspective, the percentages employed for the chi square contrast were the following: 50.08% of Spanish franchisees operate in retail, 6.90% operate in personal services, 26.11% are dedicated to specialized services and the rest operate in restoration. Results showed a chi-square statistic with a value of 1.717 (df = 2) related to a p-value of 0.424.

The respecified model for franchisor companies was necessary to assure its discriminant validity. The respecification was carried out by theoretical arguments and consisted on considering trust as a global scale on the one hand (Morgan & Hunt, 1994; Geykens & Steenkamp, 1995; Kumar, Scheer & Steenkamp, 1995a,b) and it was necessary to include the item LONG1 (“The relationship with the franchisees can be defined as a long term alliance”) in the long-term orientation construct (Anderson & Weitz, 1992) on the other hand, since it refers both the temporal facet of commitment as well as the expectation of continuity (Lusch & Brown, 1996).

Given the large number of constructs and measures employed, the theory on confirmatory factor analysis recommends an analysis of submodels (Bentler & Chou, 1987).

Although a high correlation between trust and satisfaction exists, there is evidence of discriminant validity, since the majority of constructs pass both the confidence interval and the χ ² difference test and all constructs pass at least one of these tests of discriminant validity (Garbarino & Jonhson, 1999).

For franchisee's trust all the second order factor loadings are high and significant (Credibility: γ ₁  = 0.920; Benevolence: γ ₂  = 0.920).

See also Anderson & Narus (1990); Anderson & Weitz (1992); Ganesan (1994) and Garbarino & Johnson (1999).

Fit Statistics for Rival model from franchisors side: χ 2 Satorra = 310.1267; p-value = 0.0000; SNCP = 1.608; GFI = 0.783; RMSR = 0.199; RMSEA = 0.093; ECVI = 4.416, NNFI = 0.841; IFI = 0.867; CFI = 0.864; PNFI = 0.648; Normed χ ²  = 1.911.

Brown et al. (1995) propose that the greater normative commitment is (identification and internalization), the higher performance will be, whereas instrumental commitment (conformity) reduces performance. As for normative commitment, they obtained positive parameters in the total sample and in the sub-samples in which a party of the relationship is more powerful than the other; however, when both parties have the same power, the relation between normative commitment and performance is negative.

Fit Statistics for Rival model from franchisees side: χ ² Satorra = 506.0049; p-value = 0.0000; SNCP = 6.199; GFI = 0.612; RMSR = 0.350; RMSEA = 0.147; ECVI = 10.329, NNFI = 0.743; IFI = 0.772; CFI = 0.770; CFI Robust = 0.896, PNFI = 0.671; Normed χ ²  = 3.172.