Industrial diversity in patterns of productivity convergence and trade specialisation

Trade specialisation according to current comparative advantage need not necessarily be an optimal development strategy in a long-run dynamic setting. Given the possibility of aquiring omparative advantages in industries where a country is not competitive initially (i.e. through policy intervention, learning by doing, knowledge and/or technology spillovers), we argue that ountries should try to enter high tech industries as early as possible. Based on the presence of stronger intra-industry versus weaker inter-industry spillovers, we want to test the hypothesis that countries which specialise in the lower end of the medium-high tech activites are rewarded by faster productivity increases also in the upper end of the high-tech industries. In contrast, early specialisation in less technology intensive branches yields positive spillovers mainly inside these sectors resulting in a lower long-run growth potential. We present econometric results including four industry segments for 37 countries over the time period 1981 to 1998. Our estimations confirm the hypothesis of an 'advantage of backwardness' at the branch level. Further, intra-branch spillovers were often significant while only few inter-branch effects were found. With respect to trade integration, the results were mixed, openness on the import side had positive effects in general, while openness on the export side spurred convergence only in less technology intensive industries. Besides these direct effects, only very few trade related inter-industry effects were found.