Railroad development, temporal-spatial externalities, and growth spillover: Theory and empirical evidence

This paper develops an extended Solow model by considering the time-variant spillover effect of interregional economies. Instead of using fixed spatial distances among regions, the regions' temporal-spatial distances are measured by constructing a database of train types and calculating the fastest region-to-region routes. We collect provincial economic data and railroads data over the period 2003-2012 in China to establish a dynamic temporal-spatial panel model. The estimation results indicate significant temporal-spatial externalities. Compared with classical spatial models, the temporal-spatial model not only has a much better goodness-of-fit but also has consistent total effects, direct effects and indirect effects.