The impact of convergence in the industrial mix on regional comparative growth: Empirical evidence from the French case Fre´ de´ ric Carluer 1 , Guillaume Gaulier 2 1 PEPSE, University of Grenoble II and MATISSE, University of Sorbonne, 1241 Rue des re´sidences 38040, Grenoble Ce´dex 9, France (e-mail: carluer@univ-paris1.fr) 2 CEPII and EUREQua, University of Sorbonne, 9 Rue Georges Pitard, 75015 Paris, France Received: November 2000/Accepted: February 2004 Abstract. The aim of this article is to explain heterogeneities in French re- gional labor productivities since the mid-seventies at both aggregate and sectoral level. This paper extends the works of Baumol and of Barro and Sala- i-Martin, firstly by pointing out sources of growth linked to the new growth theories (research effort, size effects) and secondly by emphasizing the impact of cross-sectoral labor reallocations through a shift-share analysis. Our re- sults show the importance of regional asymmetries and the key role played by the dynamic of sectoral composition in the convergence of labor productiv- ities within France. JEL classification: R11, O4 1. Introduction Until recently, the empirical studies of growth and convergence initiated by Baumol (1986) and Barro and Sala-i-Martin (1991) paid little attention to structural aspects, and especially to the sectoral dimension of economies, with analyses being based rather on aggregate data. Dollar and Wolff (1994) and Bernard and Jones (1996 a,b) are early exceptions at a national level. Recent papers drawing on regional data have emphasized the importance of a dis- aggregate analysis at sectoral level (Paci and Pigliaru 1997; Cuadrado-Roura et al 2000; Esteban 2000; Sorenson 2001; De la Fuente 2002; Tumpel- Gugerell and Mooslechner 2003; Carluer 2004). They suggest that to take the continuous process of sectoral reallocation of resources into account could be decisive in understanding growth and convergence. Moreover, the regional level constitutes an appropriate framework for the study of convergence, from both the empirical and theoretical points of view. Empirically, it is easier to compare data derived from the same sources than to undertake international comparisons. Theoretically, the assumptions such as those regarding the homogeneity of infrastructures, as well as the Ann Reg Sci (2005) 39:85–105 DOI: 10.1007/s00168-004-0201-1