9
which occurred despite coordinated campaigns to increase the number of women in
elective office, provided further fuel to quota fires.
The percentage of women actually elected provides an important measure of
inequitable representation, but it does not allow us to distinguish gender bias in the
electorate from gender bias in the parties. The percentage of women nominated as
candidates provides a more accurate measure of the way in which gender bias at the party
level affects election outcomes. My analysis of data on candidate gender was facilitated
by the fact that it is easy to identify male vs. female names in Mexico, and in Spanish-
speaking countries generally.
4
Figure 2 presents data on women nominated to PR lists in
each of the past 5 elections for Mexican Chamber of Deputies. It shows that female
candidates reached the 30% threshold on only three occasions before 2003: the PRD in
1997 and 2000, and the PRI in 2000.
5
[Insert Figure 2 here]
The level of bias against women increases if we examine the percentage of women
nominated to electable positions, generally considered to be the top ten positions in each
of the five PR lists. Figure 3 shows the percentage of women nominated to the top 10
spots on the PR lists over the past five elections for the Mexican Chamber of Deputies.
Women were 30% of the electable candidates only for the PRD in 1997.
[Insert Figure 3 about here]
Variation across the parties in these elections reflects the impact of variation in internal
party rules about the nomination of women (Bruhn 2003). These data suggest that gender
bias occurs prior to the election itself, among those who select candidates rather than
voters per se.
Convention