Convergence or Twin Peaks: The Spanish Case

Abstract

This paper aims to explore the Spanish geographical dynamics of per capita income among provinces, during the period 1955-1997. First, we obtain some preliminary evidence of their behavior over time from the visual inspection of their non-parametric density distributions. These distributions turn to be unimodal in virtually all cases. Next, we compute the transition matrices and the long run equilibria by means of the Markov Chain's approach, (along the lines of the contributions of Quah). Basic outcomes of this analysis also tend to predict a patter of convergence among provinces for the period 1955-97. Therefore, and according to these results, the existence of twin peaks in Spanish income distribution does not seem to be supported by the empirical evidence presented here.

Other tentative conclusions of the paper are the high level of persistence in the relative position of provinces, consistent with a low degree of mobility in the income distribution. Notwithstanding this fact, the paper detects some particular changes in the placement of provinces according to their per capita income. More specifically, the richest provinces tend to concentrate gradually in the North-East of the country, which may be attributed, in turn, to externalities linked to localization or to the proximity to the rest of Europe. Sectoral reallocation to the industrial sector also seems to have fostered growth of per capita income in those provinces that have experienced it over time.

Keywords: convergence, Twin peaks, Markov chains

JEL Classification: O4, O5, R1