Theses Doctoral

Origins and Use of Presidential Polling in Mexico, Presidential Approval in Mexico, Government Spending and Public Opinion in Mexico

Torres-Reyna, Oscar

This three-paper dissertation aims to contribute to the study of the Mexican presidency, in particular, to the understanding of the origins and use of presidential polling, its role in the policy activity of the president, and the dynamics of presidential approval between 1989 and 2011. The dissertation draws upon the presidential polling, opinion-policy and approval research done in the United States. The first paper explores a topic that has not received much attention in Mexico, the origins and use of the presidential polling unit (PPU). The second paper focuses on presidential approval in Mexico, and the third analyzes, yet another understudied topic, the relationship between government spending (used as proxy for policy) and public opinion (collected by the PPU). The first paper relies on crosstabulations, text analysis, wordclouds and cluster analysis. Additionally, to offer an insider's view, I conducted a series of interviews to seven presidential staffers during the administrations of Presidents Carlos Salinas de Gortari (Dec/1988-Nov/1994), Ernesto Zedillo Ponce de León (Dec/1994-Nov/2000), and Vicente Fox Quezada (Dec/2000-Nov/2006). The second and third papers made use of vector autoregression models to account for feedback effects among the spending and opinion variables, controlling, at the same time, for a possible `backwards' process in the opinion variables. The main assumption is that the variables are connected: all variables depend and/or explain each other.The first paper entitled "Origins and Use of Presidential Polling in Mexico" addresses the questions of what caused the creation of a government office dedicated to gauge public opinion, what poll information the presidents collected, and how it was used. I will argue that the institutionalization of public opinion within the presidency responded to the dynamics of the political system, in particular, to the changes in the electoral system and the outcome of the presidential election of 1988. The election of 1988 changed Mexico's electoral map and reconfigured the party loyalties against the ruling party PRI. Aware of this new political context, President Salinas used polling not only to study the political behavior of the Mexican voters but also as an alternative to verify electoral results. In fact, the first mandate of the presidential polling unit was to track political preferences. Eventually the use of public opinion polls expanded to other issues and became part of the presidential policy toolkit. As Jacobs and Shapiro (1995) pointed out in the case of the Kennedy administration, the Mexican presidency had now an office with "routinized procedures" to research and collect public opinion data. To identify the type of polling information collected by the presidents, in addition to interviews to presidential staffers, I applied text analysis on titles of all presidential polls conducted between 1989 and 2006. While all presidents collected opinion data on their approval ratings and customized their polling operations according to their own policy agenda, there were some overall differences. President Salinas centered his field polling operations around policy, and his phone polls for elections and presidential image. President Zedillo used field polls mostly for electoral issues and phone polls for image and communications. President Fox focused the field polls for government evaluation and customer satisfaction, and his phone polls for image and evaluation of political figures. How public opinion information was used remains an open chapter. All presidential insiders mentioned that information from public opinion polls was not specifically used to design policy but rather to test it, and to see what worked and what did not work. Polling was used to find ways to convince the public of the benefits of the presidential policies and actions. From this analysis, the conclusion is similar to what Jacobs (1992) argued in his paper on recoil effect. The presidents did use polling to try to move public opinion to their side, but also polling was used to understand what was in the mind of the public. Eventually, these efforts, I believe, made a significant contribution to the development of political public opinion and, most importantly, to the development of democratic values among the political elites.The second paper entitled "Presidential Approval in Mexico" looks at the factors that influence presidential approval using as reference research done in the United States and Mexico. I am looking for evidence that presidential approval in Mexico depends on factors directly connected to policy outcomes (Erikson, MacKuen and Stimson 2002). The risk of manipulation is at the center of this connection. The president may create the illusion of meeting the public's expectations (Kernel 1997) and/or opinion elites may misled the public against the president (MacKuen, Erikson and Stimson 1992). The argument here is that as long as presidential popularity is rooted in objective measures related to policy or economic outcomes, approval may actually be a reliable indicator of citizen's response to government actions and, therefore, a reliable measure of the president's political capital. Thus, the research question is whether approval depends on objective measures of the economy (and the overall situation of the country) or relies on the public's perceptions about the current conditions of the country. Furthermore, are those perceptions retrospective or prospective? Do they rely on what has been done or what is expected to be done? The findings presented in this paper confirm the expectations that the popularity of the Mexican president depends mostly on how the economy is doing and how the president deals with current salient issues like public safety (Buendia 1996; Gómez-Vilchis 2012). At the level of perceptions, prospective evaluations of personal well-being have a positive impact on approval but only among the richer segments of the population. It is important to notice that these perceptions are strongly influenced by the unemployment rates. The overall conclusion is that presidential approval in Mexico is rooted in macroeconomic, salient and subjective measures that are also connected to the dynamics of leading economic indicators. Presidential approval in Mexico depends, so far, on the president's capacity to solve problems.The third paper entitled "Government spending and public opinion in Mexico" explores the relationship between policy and public opinion. While this paper draws upon the opinion-policy research done in the United States, it departs from the policy preference approach to a perspective centered on policy outcomes. The main opinion variables included in the models refer to retrospective and prospective evaluations of personal well-being. These are generic and, in the question wording, do not refer to any issue in particular. One of the goals is to find whether these opinion variables are directly connected to trends in leading economic indicators (like growth of GDP percapita, unemployment, inflation). If such connection exists, then they may represent citizen's responses to current state of affairs of which the president and the government in general are perceived as responsible. This is, the opinion variables can be taken as responding to policy outcomes. The main underlying logic follows the Mood and Thermostatic models suggested by Erikson, MacKuen and Stimson (2002) and Soroka and Welzien (2010) respectively. If people started to feel that things are getting worse, then I would expect the government to increase spending, for example to stimulate the economy. Conversely, if people feel things are getting better, then I would expect the president to scale back on spending. The models show feedback in the economic but not in the public safety models (this is, the reciprocal effect between opinion and spending). In the models where economic spending is the contemporaneous outcome variable, positive prospective evaluations of personal well-being and perceptions that the economy is the most important problem (MIP) facing the nation show significant effects on spending. In the case of spending on public safety, negative prospective evaluation of personal well-being and the perceptions that public safety is the most important problem in the country play a significant role (but there is no feedback). An important finding is that the public attentiveness to economic issues (MIP) does explain a significant portion of the variance in spending on the economy. Regarding the impact of opinions by socioeconomic status, there is not enough evidence to conclude that the President listens more to a particular segment of the population. The results, however, seem to indicate a marginal difference in favor of the public with lower income and education levels. Overall, the findings presented here show a connection between presidential spending activity and public opinion. This suggests some responsiveness towards public opinion. Regardless of their own personal agendas, presidents have worked to improve the conditions of the citizens and responded to their perceptions of the general situation of the country. The fact that most of the population is still poor combined with the fact that polling is here to stay (along with the new impact of social media), has forced politicians to be responsive to the needs and wants of the public. As long as the public remains connected to its economic reality and pay attention to their immediate environment, any attempt of manipulation will not last long. The Mexican public is wise and, repeatedly in electoral processes, it has demonstrated strong and reasonable political culture. Mexican politicians are catching up with the public and this is a good thing. However, as democracy consolidates in Mexico, it may be possible to see the nature of responsiveness changing as the influence of traditional political elites fades and other forms of influence start taking over. Mexico is still in a democratic honeymoon.

Geographic Areas


  • thumnail for TorresReyna_columbia_0054D_11632.pdf TorresReyna_columbia_0054D_11632.pdf application/pdf 5.15 MB Download File

More About This Work

Academic Units
Political Science
Thesis Advisors
Shapiro, Robert Y.
Ph.D., Columbia University
Published Here
October 25, 2013