The Legacy of History - The Political Geography of Inequality - page 210

TABLE 8.1. Dimensions of the Geography of Income Inequality, 1980-2000

OLS Pcse

Robust Pcse (ar1)

β β β

(s.e) (s.e) (s.e)

Interregional Differences 2.59*** 2.59*** 1.44

in Growth (.81) (.81) (1.2)

Interregional Differences in the .312** .312** .17

Incidence of Unemployment (.12) (.12) (.20)

Trade − .0025*** − .0025*** − .0024***

Openness (.0006) (.0006) (.0008)

Mobility − .0058* − .0058* − .0009

(.003) (.003) (.0008)

Intercept 1.07*** 1.07*** 1.23***

(.10) (.10) (.131)

N 60 60 60

F/Wald Chi 11.39 346.28 141.34

R-squared .47 .62 .76

Key: * p < .10,**p < .05, *** p < .01

regions, and reduce the differences in the shape of the income distribution
between regions, thereby reducing incentives to pursue or sustain separate tax
rates. 2 These results offer direct support for the hypothesis that mobility reduces
interregional differences in economic geography, working to facilitate the cen-
tralization of interpersonal redistribution (hypothesis 1). The next hurdle is to
assess how these inequalities, as proxy for heterogeneity of preferences between
regions, are reflected in different choices about territorial fragmentation of the
system of interpersonal redistribution.

Inequality, Representation, and Fiscal Structures

Concerning the translation of preferences into outcomes, the central claim
throughout this topic is that the way contending preferences for fiscal struc-
tures actually translate into observable levels of interpersonal redistribution is
contingent on the balance of power between the center and the subnational

2

It should be noted, however, that the variable capturing exposure to external shocks (trade
openness) is the only one that shows consistent effects across the different specifications. Once a
potential process of serial correlation is taken into account, the effects of all other determinants
are reduced by half and cease to be significant, although they consistently show the right sign.
As a result, trade openness appears to be the only robust “good instrument” in equation (8.1) .
In contrast, the proxy indicators for income and risks differences are good instruments, yet
not robust to autocorrelation. Finally, mobility is a weak instrument throughout the different
specifications.