Geoscience Reference
In-Depth Information
For reasons related to the agglomeration of production (Krugman
1991
),
industrial production tends to cluster in particular regions within any given
union. As a result, in line with the previous estimates of (8.1), the larger the
pattern of interindustry labor mobility, the larger the scope of cross-regional
externalities within the labor force, and other things being equal, the more
homogenous the wage structures across industrial regions. By implication, large
levels of interindustry labor mobility serve as a proxy for interregional mobil-
ity in that they work to even the geography of inequality among subnational
units.
11
As in the previous section, this provides an indirect test for the connec-
tion between mobility and preferences for the centralization of interpersonal
redistribution posited in hypothesis 1.
Cross-regional differences in production also relate to the importance of oil
dependency. Those unions with a lower dependency on oil imports are likely
to have more heterogeneous regional economic structures, leading to larger
interregional income differences (Wibbels and Goldberg
2010
). In addition,
a recent line of research has established that the territorial distribution of
different ethnic groups is endogenous to physical geography (Michaelopoulos
2008
). As a result, one would expect unions with larger levels of regional
ethnic concentration to have a more uneven geography of income inequality.
The results are presented in
Table 8.3
.
The results offer a mixed picture on the instruments selected. The expecta-
tions concerning the concentration of ethnic groups do not receive any empir-
ical support. In turn, the notion that oil dependency is negatively correlated
with interregional income disparities is only consistent with some of the results
reported and lacks robustness across the different specifications. In contrast, the
proxy for interregional mobility is negatively and consistently associated with
interregional income differences. This association provides additional empirical
grounds for the hypothesis that mobility drives the centralization of interper-
sonal redistribution indirectly through its effect on the geography of income
inequality (hypothesis 1).
12
11
The measure of interindustry labor mobility comes from Zhou (
2009
) and is defined as the
elasticity of labor to interindustry wage differentials. The rationale behind the measure is
defined as follows by Zhou (
2009
): “The advantage of elasticity measure over either the real
labor movement or the variation of interindustry wage differentials is that it takes into account
both employment shifts in the labor market and the factor price (i.e., wages) changes that
purport to incentivize the employment shifts. In other words, it looks at the sensitivity of
rational labor owners to the observable wage differentials across industries [ ...]. Ideally,
higher degrees of the sensitivity of labor to factor price differentials would mean higher levels
of ILM, and vice versa for lower degrees of sensitivity. Therefore, empirically we should be able
to distinguish economies that have higher labor mobility from economies that have lower labor
mobility by finding higher degrees of labor-wage sensitivity in the former and lower degrees of
labor-wage sensitivity in the latter set of economies.”
12
These findings are robust to the inclusion in equation
(8.1)
of all the other instruments used to
predict, in equation
(8.2)
, the centralization of interpersonal redistribution (reported in
Table
8.4
).
