26
The question is why legislative fragmentation would lead to more reform? Since
my theory predicts either change or stability with a low level of institutional or partisan
constraint (in this case a cohesive legislature), cohesion should lead to reforms only when
the legislative majority prefers reform to the status quo. With a low level of cohesion, any
change should be difficult to negotiate and implement. This means that legislative
cohesion should have a positive effect on structural reform, but that the significance of
the variable may be lower compared to a situation without an association of the variance
of reform and cohesion. As the statistical analysis shows, however, cohesion has a
negative effect. Thus, a high level of partisan differences in the legislature is more likely
to promote reforms than a low level of differences is.
In fact, since I am attempting to control for preferences by assuming that
decision-makers should prefer reform to status quo (or at least anti-reform legislation),
even parties in fragmented legislatures should be able to unite and implement reforms.
Thus, despite partisan differences on other issues, they should be able to negotiate reform
changes. But why is a fragmented legislature more likely to enact reforms than a system
with a high level of cohesion (a low number of effective parties) in the legislature?
According to my theory and assumptions these parties should also prefer reform.
One explanation is concerned with how the dependent variable is operationalized
and the observations included in the analysis. If one looks at the rate of reform in
different countries, one will notice that reform during 1980-1995 have taken place in
countries that were at an already low level of reform. That is to say, compared to other
Latin American countries, certain countries have high levels of external tariffs, state-
owned enterprises or high controls of capital. In these countries there is a greater need for
Convention