The Effect of Franchising on Firm Growth: An Empirical Analysis from China

Abstract. This study examines the effect of franchising on firm growth in China’s franchise systems. The methodology involves estimating growth model using unbalanced panel data compiled from the China Top 100 Franchises from 2008 to 2011. The results indicate that franchising has marginally significant positive effect on firm growth. A related significant funding is that inner firm conditions such as preceding firm size and average sale of units have negative effect on firm growth.

Key words: China, firm growth, panel data, effect

1. Introduction

In the late 1980s, the word “franchise” was introduced to China. Experienced three stages of introduction, adoption and rapid development, franchising plays a prominent role in China’s business life today (Zhiqiong June, Zhu, & Terry, 2008). For example, in 2011, the number of franchise systems broke 5,000 that up 25% over 2008, and the total number of franchised stores exceeded 1 million. By industry, franchise systems covered more than 70 industry formats in which mainly concentrated retailing, catering, and service industries. A nature question is, when firm emphasizes franchising in its expansion, what on earth the effect on its growth?

Past research has studied the relationship between franchising and firm growth in Western (Elango & Fried, 1997); Gillis & Castrogiovanni, 2012). However, the role of franchising in stimulating firm growth is a controversial issue in the franchising literature. Lafontaine (1992) found a positive relationship between franchising and firm growth confirming the resource scarcity theory ((Oxenfeldt & Kelly, 1968); Norton, 1988), while Michael (2000) found a negative relationship. Alon (2001) found that larger franchise systems franchise more, while Castrogiovanni, Combs, and Justis (2006) found that the bigger the franchisor the less its propensity to franchise.

To gain a clearer picture of franchise development in China, this paper represents an empirical analysis to shed light on the effect of franchising on firm growth, and investigated the relationship between franchising and firm growth in China.

2. Data

The data set refers to an unbalance panel of 181 firms observed from 2008 to 2011, with 460 observations. The main data source is china Annual Franchise 100 published from 2008 to 2011 by China Chain Store & Franchise Association’s (CCFA) annual survey filling by China’s franchisors. The data set contains information in various firms comes from various sectors, not only limit to one sector (e.g. restaurant sector) in previous studies, which allows us to revisit the issue of firm growth and franchising in a much more efficient and precise way. It should be pointed out that the number of franchisors included each year is rather unstable, those firms that appear are not all the same from one year to the next. Table1 gives descriptive statistics for the main variables of 460 observations.

3. Model

The empirical analysis used in this article is based on the conditional convergence model that is commonly used in the literature (Sen, 1998; Dant, Kacker, Coughlan, & Emerson, 2007). Specially, the following standard growth regression model is estimated to test the hypothesis:

Where ηi is the unobservable firm specific effect, and εit is the error term.

Model (1) has following assumptions: 1) Linearity, that the model is linear in parameters βk, k=1

3.1 Independence, that the observations are independent across firms but not necessarily across time; 3) Strict erogeneity, that the idiosyncratic error term εit is assumed uncorrelated with the explanatory variables of all past, current and future time periods of the same firm, and with the individual specific effect ηi ; 4) Homoscedastic and no serial correlation in error variance; 5) Fixed effects and random effects, if ηi is correlated with the explanatory variables the model is fixed effects model; if ηi is not correlated with the explanatory variables the model is random effects model, 6) Polling effects, that if ignore the panel structure and get the ordinary least square linear regression model. We use the growth in the number of total outlets to assess the firm growth, and use the growth in the number of franchised outlets to assess the use of franchising In model (1). In equation (2), we neglect the time specific effect. We discuss briefly the properties of ui,t below; more details see (Hisao, 2003; Green, 2003; Wooldridge, 2002).

4. Methodology and Estimation

In order to evaluate the effect of franchising on firm growth, an analysis was carried on a data set compiled from the annual China Franchise Top 100 (2008-2011), and subsequently analyzed using the R 2.14 statistical software program and the plm package (Croissant & Millo, 2008). In our econometrical analysis, the log-log form is chosen to make residuals are as far as possible to approximate a normal distribution.

Under the assumption that ui,t is the same across firms, equation (1) can be estimated by OLS on a cross section of firms. The problem of pooling effect is that it neglects an unobservable individual effect which leads to firm specific heterogeneity. Such unobservable individual effect is potentially correlated with explanatory variables. If it is not controlled for in the estimation, the parameters estimates will be inconsistent due to omitted variable bias. In order to avoid this problem, we use a panel model method. The panel method allows controlling for individual effects. It also allows the use of more observations and gives more degrees so freedom. The usual F-test is used to select the most appropriate model between the pooled cross section effect and the fixed effect models. Hausman’s test is also computed to compare fixed effect and random effect models.

In this paper we estimate the effect of franchising after controlling for other determinants and potential bias induced by the firm specific effects. In Table 2, we present pooling effects, fixed effects and random effects estimates. Empirical results of the pooling effects model show that the effect of franchising on firm growth is positive and statistically significant at the 0.01 level. All control variables, lagged total outlets and average sales of units, included in the model also have the expected signs. However, lagged total outlets are not significant at 0.10 levels. The results of F-test indicate that the pooling effects model is not accepted.

In the random effects model, franchising has its expected (positive) sign and enters the firm growth regression significantly at 0.10 percent level. The results also show that all the other variables have the expected sign. Lagged total outlets are not significant at 0.10 levels; meanwhile, average sales of units are significant at 0.10 levels. The Hausman’s test indicates that the random effect model is not preferable to the fixed effect model.

In the fixed effect model, all variables have the expected signs and statistically significant at 0.1 level. The results of F-test indicate that the fixed effect model is preferable to the pooling effect model. Although the measure of goodness of fit, R2 = 0.57, is less than the pooling effect model (0.57) and the random model (0.64).

Fig 1 shows the fixed effects of 181 firms. From Fig 1, we can tell that the firm growth rates of firms located in the interval (-1, 1) with some outliers located far in the right interval. The average of fixed effects is -0.21, the median of fixed effects is -0.26. There are one firm has very high growth rate 2.93. Cancelling the outlier (which is greater than 2), the mean and median are -0.24 and -0.27 respectively. On the whole, the growth rate of China’s franchise systems is decrease not increase.

5. Acknowledt

This article establishes, through an analysis of firm level data, that there is a close relationship between the use of franchising and firm growth within China’s franchise systems. This is predicted to occur because franchising helps firms overcome various constraints that inhibit growth, by providing a bundle of financial capital and managerial talent. The paper also suggests that a firm’s growth is likely to be influenced by its past growth strategy, especially by its last period firm size and current period average sale of units.

Our findings indicate that the effect of franchising on firm growth is positive but statistically significant at 0.01 levels which confirm the resource scarcity theory. This result is similar to those found in Lafontaine (1992) and Sen (1998). Our results indicate that the effect of last period firm size and average sales of units is negative and statistically significant at 0.01 level, which confirming the found in Castrogiovanni, Combs, and Justis (2006).

The finding that franchising has the significant and positive effect on firm growth has an important implication for China’s franchise chains. If the firm increase hundred percent of the outlets, it is likely that the firm has expanded itself about thirty percent. This result show that Why more and more China’s firm adds outlets and increased its use of franchising to achieve firm growth. The effect of franchising on firm growth accompany with an old question of whether there are limits to the use of franchising as a growth strategy. The study show that the benefits provided by company owned outlets will have negative effect if the chain has too high rate of firm growth in previous period and/or larger amount of sales. Thus, the firm is likely to be influenced by its past growth path.

References

[1] Zhiqiong June W, Zhu M, Terry A. The Development of Franchising in China. Journal of Marketing Channels. 2008;15(2/3):167-84.

[2] Elango B, Fried VH. Franchising Research: A Literature Review and Synthesis. Journal of Small Business Management. 1997;35(3):68-81.

[3] Gillis W, Castrogiovanni GJ. The franchising business model: An entrepreneurial growth alternative. International Entrepreneurship and Management Journal. 2012;8(1):75-98.

[4] Lafontaine F. Agency theory and franchising: some empirical results. RAND Journal of Economics (RAND Journal of Economics). 1992;23(2):263-83.

[5] Oxenfeldt MR, Kelly AO. Will Successful Franchise Systems Ultimately Become Wholly-Owned Chains? Journal of Retailing. 1968;44(4):69.

[6] Norton SW. An Empirical Look at Franchising as an Organizational Form. Journal of Business. 1988;61(2):197-218.

[7] Michael SC. Investments to create bargaining power: The case of franchising. Strategic Management Journal. 2000;21(4):497-514.

[8] Alon I. The Use of Franchising by U.S.-Based Retailers. Journal of Small Business Management. 2001;39(2):111-22.

[9] Castrogiovanni GJ, Combs JG, Justis RT. Resource Scarcity and Agency Theory Predictions Concerning the Continued Use of Franchising in Multi-outlet Networks. Journal of Small Business Management. 2006;44(1):27-44.

[10] Sen KC. The use of franchising as a growth strategy by US restaurant franchisors. Journal of Consumer Marketing. 1998;15(4):320-407.

[11] Dant RP, Kacker M, Coughlan AT, Emerson J. A cointegration analysis of the correlates of performance in franchised channels. 2007. p. 169-89.

[12] Croissant Y, Millo G. Panel data econometrics in R: The plm package. Journal of Statistical Software. 2008;27(2):1-43.