Geoscience Reference
In-Depth Information
The control variables display the anticipated effects. Market inequality is
obviously a strong and positive predictor of disposable income inequality.
In addition, the indicators of economic coordination and partisanship confirm
standard results in comparative political economy. Organized capitalism and
left-wing incumbency contribute, even if marginally, to a more egalitarian dis-
tribution of resources in society (Beramendi and Anderson
2008
; Iversen
2005
;
Iversen and Soskice
2009
; Kenworthy
2004
; Pontusson
2005
; Rueda
2008
;
Stasavage and Scheve
2009
).
Turning to the central concerns in this topic,
Table 8.5
indicates a strong and
positive effect of the level of decentralization of interpersonal redistribution on
the observable amount of disposable income inequality. On the basis of the
estimates reported in the first column, a country with a fully centralized system
of interpersonal redistribution is predicted to have a level of overall disposable
income inequality of 0.21. In contrast, a country with a territorially fragmented
system of taxes and transfers is predicted to have a level of inequality of 0.34.In
other words, the territorial organization of public insurance systems accounts
for a 30% change in the spread of disposable income, even after controlling
for all the standard factors in the literature.
On the basis of this result, it would seem tempting to rescue the conven-
tional view (
Chapter 1
) and elegantly proclaim that decentralization
creates
inequality. A central theme in this study is that such an impulse is misleading in
that the conventional view is, at best, incomplete. And this is so because decen-
tralization itself is a function of a politico-economic process determined by the
interplay between economic geography and political representation. This being
the case, the indicator of decentralization of redistribution and the interaction
between geography and representation ought to work as substitutes of one
another as predictors of the spread of disposable income in society. A compar-
ison between the two columns in
Table 8.5
reveals that this in fact the case.
Put simply, the variables capturing the generating process of decentralization
itself perform as well in predicting the overall levels of inequality, holding all
the other factors driving the latter constant.
This is best illustrated by analyzing the conditional effects of interregional
inequalities on the overall level of inequality (
Figure 8.7
). The core result is
clear: the strength of the association between interregional and interpersonal
inequality is mediated by political representation. Much like the case of decen-
tralization, higher values of representation mute the link between interregional
and interpersonal inequalities. The same conditions that drive the selection of
decentralized public insurance systems are associated with larger levels of over-
all disposable income inequality. Under centrifugal representation, territorial
inequalities sustain themselves through their influence in the selection of frag-
mented welfare systems that, in turn, work to sustain, perpetuate or reproduce
inequality. Thus, to the extent that there is an association between inequality
one and a panel tobit approach to account for the fact that the dependent variable is left-
censored.
