64
Hypothesis 4 (ideological distance to the majority party) finds mixed results
across models. The coefficient is negative and statistically significant for the clustered SE
model, while it does not have any effect for the multilevel estimation. As a consequence,
I do not find even a close support for my hypothesis, given that I expected a higher
probability of strategic behavior as distance from majority party increased. A possible
explanation may have to do with the fact that parties lying farther away from majority
party tend to have weak territorial linkages (i.e. third forces from urban districts), and
therefore do not concentrate so much time and effort in targeting local voters. However,
further research will address this counter-intuitive point. The explanation of the
differences across models may have to do with deep province-level effects that are
affecting legislators with the same territorial origin together. In fact, taking a look at the
reports of the random intercepts, they explain a substantive part of the variance of the
model. As it becomes evident in Table 5.8, variation is the norm in terms of province-
level effects. A very substantive finding is the strong and negative coefficient of the City
of Buenos Aires. Given that it has no municipal structure yet, the expectation of a
negative direction was clear, and it is confirmed now.
Looking at the coefficients for committee chairmanship, hypothesis 5 seems to be
confirmed. Committee chairmen notoriously decrease the use of local legislation, also
with a strong statistical significance. Everything else equal, committee chairs that have a
municipal background reduce the chances of sending local bills by 21% compared to
those that do not lead a COmmittee- Similarly, deputies without a past as mayors but
leading a committee are 26% less likely to submit local bills than those colleagues who
are not a chair.