Presidents play a central role in legislative activity in Latin America. Previous research highlights that some form of ideological compromise on behalf of the president is vital to sustain successful legislative coalitions. Yet, primarily due to the lack of a firm empirical basis on which to measure such presidential give-and-take, the extent to which presidents make use of such policy compromise, and under what conditions this is a viable strategy, remains unknown.
One of the primary obstacles towards a better understanding of these dynamics has been – to date – the difficulty of deriving reliable comparable estimates of the policy compromise of presidents over time. We collected ‘State of the Union’ addresses for 73 Latin American presidents between 1980 and 2014 and we used the Wordfish algorithm (Slapin and Proksch 2008) to provide a position for each one of these presidents, for each year, on the main latent political dimension. The heatmap below illustrates these positions. Each country’s time series reports standardized z-scores and progresses from the earliest speech available in our sample on the left to the most recent observation on the right. The shading of the cells reflect the estimated ideal-point; more negative values are shown in darker shading, increasingly becoming lighter with more positive values. Note that absolute positions can only be compared within a country. Cross country comparisons are only possible for presidential movements, because standardisation per country expresses movements relative to the stretching of the respective main political dimension.
Executives in Colombia, Costa Rica, Uruguay and also Venezuela all display consistent and stable political views. In contrast, in Argentina, Chile and Mexico, the presidents’ movements appear to trend over time. The remaining countries demonstrate a high mobility of presidents along the latent issue dimension.
With this cross-national time-series, we can explore to what extent presidents engage in policy compromise, and under what conditions this is a viable strategy. Our central finding is relatively straightforward. We show that the president does not adopt a static policy position across her term. Rather, the president, if she wishes to pursue a statutory legislative agenda, will respond to shifting dynamics in the house. Specifically, we suggest that when the position of the median party changes, the president will shift her policy position in the same direction. Of course, the degree to which a president is willing to compromise will depend on a number of conditioning variables; specifically, the president’s non-statutory power, her government status and her ability to offset the need for compromise with increased material transfers (see the figure below). At high levels of executive power, and when the president has access to large amounts of discretionary funds to use as pork, she will compromise less in response to changes in the median party position. We also demonstrate that when the president compromises her position in response to the median party in this manner, a president will enjoy a higher rate of success for her legislative initiatives than were she not to do so.
We think our results have some important implications. Presidents in Latin America are not always the inflexible and imperial leaders as previously characterized by Juan Linz. However, they also show that under certain circumstances they can be, in particular, in minority situations. In short, institutional variation among separation of power systems will condition the degree of harmony between the executive and legislative branch. The flip side of the coin is of course that presidents are willing to compromise given the right institutional incentives. Our results show that a president can influence the interbranch relationship with signals of policy compromise. Our results on policy compromise are indicative of this inter-branch legislative dynamic that helps the president to build and maintain coalitions, and pass legislation in the house.