Scale and composition effects of human capital on Spanish regional migration

Abstract

The aim of this article is to test the effects of human capital on Spanish internal migration, decomposed into two components: the size and the composition of the labour force. Our results indicate that those Spanish regions that experienced increases in the ratio of skilled to unskilled workers attracted less immigration in the period 1965 to 1984. Between 1985 and 2000, when employment rates turned nonsignificant in determining migration, the opposite was observed.

Keywords:

• migration
• human capital
• scale effect
• composition effect

JEL Classification:

• J61
• J31
• O15
• R11

Notes

¹ There are two regions with index i = 1, 2, respectively, where 1 is the specific region and 2 the whole set of the remaining regions.

² This implies that marginal product of capital may diverge across regions (Yamamoto, 2008). We thus relax the assumption of perfect mobility of capital imposed in Reichlin and Rustichini (1998), and test it explicitly with our data.

³ Skilled are those with secondary noncompulsory education or higher, with unskilled otherwise.

⁴ The PMG, Pesaran et al. (1999) constrains the long-run coefficient vector to be equal across panels while allowing for region-specific short-run adjustment coefficients. A similar approach has been followed by Fachin (2007) for estimating internal migration in Italy.

⁵ Bover and Velilla (2005), Maza and Villaverde (2004) and Hierro (2009) concur that in the mid-1980s, there was a shift in migration behaviour due to the increase in both intra-regional flow movements and return migration, thus contributing to an increase in gross inter-regional flows, but not in net flows.