type of civic organization are likely to press for improved judicial performance. The
level of participation in elections (PART) is also used as an indicator of a politically
active citizenry. Electoral participation has been found to be important in studies of
human rights violations in Mexico (Beer and Mitchell 2004). To measure poverty the
“marginality” index is included (CONAPO 2001). State GDP per capita and the total
state population are included as controls.
[Table 1 about here]
Because of the small number of observations in the dataset and the related problems
of achieving significant coefficients in regression equations, I begin by examining the
simple pairwise correlations of all of the variables. Many of the dependent variables are
highly correlated with each other. JUSTICE, the overall indicator of justice from the
CEE survey, is significantly correlated with judicial budgets per capita and the corruption
score and overall sense of security. (The negative correlation for the corruption score is
in the expected direction since the lower the score, the lower the corruption is in the state.
Thus the greater the population’s sense that there is the correct application of justice in
the state, the lower the level of corruption in the state). Judicial budgets per capita are
also significantly correlated with the corruption score and Magaloni and Zepeda’s
performance index. The corruption score is significantly correlated with Magaloni and
Zepeda’s performance index and SECURE. Magaloni and Zepeda’s performance index
is significantly higher in states with a judicial council. The only dependent variables not
significantly correlated with any other dependent variables are the two indicators of
judicial access, human rights violations per 100,000 and the number of judges per capita.
Looking now at the relationships among the dependent and independent variables,
we find some support for the relevance of political competition. The margin of victory is
29
Convention