“…The test for differences across the predictor groups in MANOVA is based on statistics that are convertible into equivalent multivariate F ratios (Cooley & Lohnes, 1971). The procedure safeguards against the inflation in the error rate that would occur if a series (2001), Dant and Kaufmann (2003), Dant et al (2008), Lafontaine (1992), Lafontaine and Shaw (1999), Shane (1998) C2 Total network size in the domestic market Alon (2001), Dant and Kaufmann (2003), Dant et al (2008), Lafontaine and Shaw (1999), Shane (1998) C3 Average total investment required in '000s of Euros Alon (2001), Brickley and Dark (1987), Dant and Kaufmann (2003), Dant et al (2008), Lafontaine (1992) C4 Average franchise fee in '000s of Euros Dant and Kaufmann (2003), Dant et al (2008), Lafontaine (1992), Shane (1998) C5 Average ongoing royalty fee rate in percent Alon (2001), Dant and Kaufmann (2003), Dant et al (2008), Shane (1998) C6 Cash liquidity requirement in '000s of Euros Dant et al (2008), Lafontaine (1992), Shane (1998) C7 Incidence of internationalization Dant et al (2008), Perrigot and Cliquet (2005) C8 Sectoral differences (1-products and retail vs 0-services) Caves and Murphy (1976), Dant et al (2008), Lafontaine and Shaw (1999), Shane (1998), Thomas et al (1990) Data from the German Franchise Association's annual franchise guide. I further control for multiunit ownership by the average number of outlets each franchisee owns per system, with insignificant results.…”