workers to move from the high unemployment regions to the low unemployment regions
where wages are higher. This was exactly the case during the 1979-86 period when wage
bargaining was centralised.20
Second, although wage differentials have increased since 1986, the scant labour mo-
bility can be explained on the basis of the huge differentials in average housing prices
between the high and low unemployment groups of regions.
Third, the combination of rising incomes with family and government support may
have made people more sensitive to the amenities in their place of residence (de la Fuente,
1999). Attanasio and Padoa-Schioppa (1991) argue that young people, who are the bulk
of emigrants, are less willing to move when unemployed because of the support provided
by the rising family incomes. In addition, Antolin and Bover (1997) find that those
unemployed receiving unemployment benefits are less likely to migrate.
The above empirical evidence is also supported by the Eurostat database which doc-
uments very low levels of regional migration over the past decades.
Regarding firm mobility firms do not move from the low to the high unemployment
regions, where wages are lower, for the following reasons.
First, the high unemployment regions in Spain are generally peripheral and have an
inadequate endowment of public infrastructures (highways connecting poor regions with
richer ones were finished during the last decade, for instance).21 This leads to higher
transportation costs and thus limits the willingness of firms to move.
Second, in contrast to the Blanchard and Katz (1992) findings for the US, Spanish firms
do not move to lower wage regions due to aglomeration effects.22 When firms locate close
to large markets, they enjoy positive aglomeration externalities and increasing returns to
scale. Hence, moving to another region would imply an overall increase in costs (the lower
wage does not compensate for the loss of these externalities). In fact, firms have tended
to locate mainly in the richer regions of Madrid, Ebro Axis and the Mediterranean coast.
In the following sections we attempt to identify the causes of regional unemployment in
Spain by examining the interplay of labour market lags with region-specific and national
shocks in each of the high and low unemployment groups of regions.
4 Estimation Results
Our estimated dynamic structural model comprises a system of labour demand, wage
setting and labour force equations, and covers two panels of regions. A panel for the
group of the eleven high unemployment rate regions and a panel for the group of the six
low unemployment rate regions (see Table 1).
20 See Bande et al. (2005) for an intuition of the effect of centralised wage bargaining on regional
unemployment.
21 Despite the high effort by the EU to improve the infrastructure endowments of poor regions, European
regional funds have not succeeded in improving the performance of the high unemployment regions relative
to the rest of the country.
22 See Krugman (1998) for the arguments of the new economic geography on agglomeration effects.
10