Using Presidential Speeches in Latin America as Data

A few weeks ago on the pages of this blog, Phillip discussed the general trends that can be observed in the inaugural speeches of Presidents in Central and Eastern Europe. His post indicated the value of executive speeches as a potential source of comparative data. Presidents make speeches all the time and in these speeches, they often discuss policies, agendas and future legislative plans.

So I thought that this would be a good opportunity for some shameless self-promotion. With two colleagues, Christian Arnold and Nina Wiesehomeier, I have been working on a project that is using the annual addresses of Latin American presidents as data in order to derive some comparative understanding of executive politics across the region.

Every year, similar to the state-of-the-union speech in the United States, Latin American presidents make a speech to the national assembly. This is an institutionalized event, where the president is constitutionally obliged to make this speech at an appointed time each year and report on the initiatives of the executive over the last year, and the proposed agenda for the year ahead (an example of Dilma’s 2014 speech in Brazil can be found here). Although public interest in this speech varies across the region, there is one constant audience: the legislature. As such, this is a unique opportunity to comparatively explore this speech across countries, which we interpret as primarily a signal to the country’s legislators.

After a rather intensive and somewhat exhausting data collection effort, we managed to collect annual state-of-the-union addresses for 68 presidents across thirteen Latin American countries between the years 1980 and 2014 (we tried to begin with the year of the most recent wave of re-democratization in each case). We then employed computational models, based on the scaling algorithm Wordfish, to scale these speeches (relative to each other within each country).[1] We ended up with a policy position on the main dimension of political competition (i.e. whatever the major underlying issue cleavage is in that country) for each president and for each year. An example for Brazil can be found below.


Given national political and ideological contexts vary so widely across the region, we are unable to compare the absolute policy position of these presidents. We can however, compare their relative movement between one year and the next. For example, see the case of Brazil below.


This data can help us to understand the dynamics of executive-legislative relations in Latin America. We explored under what conditions presidents are more likely to signal their willingness to compromise and whether this has an effect on their rate of executive legislative success. Contrary to the somewhat pessimistic Linzian interpretation of Latin American politics (and in line with the wave of research on coalitional politics across the region), we are able to demonstrate with this data that presidents use speeches as a means to strategically signal policy concessions to coalition partners and the legislature more generally. The degree to which they are willing to do this is intrinsically linked to the dynamics of interbranch competition in Latin American presidential systems and will depend on a combination of the coalition status of the executive’s party and the legislative powers granted to the president. Latin American presidents will often compromise or alter their position in the policy space from one year to the next and this movement will subsequently have an effect on their rate of legislative success.

The upshot: these speeches are not just hot air. They are strategic signals of policy intent and compromise, which the president may use to ensure the success of their legislative agenda.

Our project website can be found here.

[1] More detail on Wordfish can be found here.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.