How to Study Leaders

Introduction

Our research was motivated by questions about the democratic quality of political leaders in Latin America. How could leaders who had successfully established democracy in their countries take political paths that would lead to a gradual deterioration of the quality of the democratic institutions? What motivates some leaders to employ populist, clientelistic methods to build or strengthen their power? Why are so many young politicians so naive and inefficient that they quickly lose power?

We observed different type of leaders in Latin America. We wanted to discover why in some countries excellent democratic leaders emerge while in others populist, inefficient or corrupt leaders prevail. The literature suggests a combination of elements linked to the historical and cultural context, the strength of state institutions and the extent of democratic practice among the citizens of the country in question. What we did not find were texts that laid out the thinking of politicians themselves. Therefore we decided to interview politicians. We wanted to know their perspective on the democratic quality of political leaders and the political conditions that help the emergence of different types of leader.

We knew, then, what we wanted to do but still had to decide how to carry out the research. Therefore we consulted both books and experts. Before we coordinated the interviews we undertook a literature review. Our first discovery was that there is very little written about leadership in academic texts, especially in Latin American political science. Although one of the classic texts on leadership, The Prince by Niccolò Machiavelli, had recognized the importance of leaders in any political project, modern political science had concentrated more on the structure of parliaments, political party systems, elections and the relations between the branches of state power, omitting the subject of leadership.

We also discovered that most of the literature on leadership was in the field of business studies, concentrating on private sector performance. Although some of the ideas in these studies can be applied to our field, we needed a different approach. In studies on business, negotiation and strategy, there are a lot of resources on both the study and training of leaders. But this material cannot really be applied to the study of political leaders. There are major differences between leading businesses and leading countries. Typically studies on leadership focus on an analysis of the personalities of the individuals. However, they tend to be more of a psychological examination of their good and bad points rather than a political discussion. Our interest was shared by other academics who had started to investigate the subject of leadership more systematically (Lord 2003;Kellerman 2004, 2012; Bueno de Mesquita and Smith 2011; Nye 2008; Helms 2012).

In order to establish a line of questioning for future interviews we carried out a series of informal meetings with colleagues in Buenos Aires and Madrid, and ran the questions by opinion poll specialists for their input. Thinking about questions for future interviewees was the first concrete task of this research. The project was taking shape. We wrote it up and submitted it to different foundations. In August 2009 the Foundation Open Society Institute in Washington, DC, approved our project and in October we started work.

The study took four years and covered five different cases, which enabled us to offer comparative analysis. It entailed ten study trips, 319 interviews in six countries, seven seminars organized in Toronto, Buenos Aires, Quito and Madrid, the writing of research reports, work documents and journalistic articles, the editing of a book in Spanish, and finally the writing of this book. It involved a team of seven research assistants in five countries.

Here we intend to discuss how we organized our material and conducted the research. This chapter is not strictly methodological. It is not always easy to implement what methodological texts stipulate. Often the researcher changes their hypothesis, variables and ideas and in many cases this means starting again from scratch. We had to grapple with social science methodological rules, but what we seek to convey stems more from concrete experience than scientific theories.

This chapter also covers the unexpected elements that any social scientist can encounter when they seek to explain political reality.

Defining our hypothesis and case studies

The initial hypothesis of this research holds that the deep-rooted authoritarian tradition in Latin America protected those in power from having to engage in any real deep and meaningful transformation of the democratic system. A consequence of this, we maintain, was the destruction of any new democratic leadership. Our assumption was that the traditional modes of behaviour such as clientelism, caudillismo, authoritarianism, the desire to hold on to power indefinitely, and corruption helped to explain the numerous political crises which have been the scourge of the region throughout the last three decades. Other countries, by contrast, worked to institutionalize political systems that broadened democracy.

Given that this is a rather sweeping assumption, we set out from the beginning to compare different cases in order to be able to reach some generalization, since, as Morlino says, 'comparison makes it possible to control for the hypothesis' (Morlino 1991: 14). In addition, a comparative study helps to construct assumptions and identify trends that we hoped would shed some light on how change could improve the democratic quality of leaders.

The examples we chose adhered to traditional comparative criteria, i.e. cases with similarities as well as differences. We used the Mill difference method (Pérez-Liñán 2007: 8–10) to identify cases in which we assumed that the leadership was democratic and others in which we held that the leadership characteristics were less democratic.

To begin with, we took a series of social phenomena which we considered to be tied to the rise of different types of leadership, as were the crises, the problems of political representation and the fragmentation of political parties. Our dependent variable was initially the democratic quality of the political leaders. As a first step in the research we defined democratic quality by taking as a basis the study by Barreda, which measures this quality as a function of five dimensions: 1) political rights and civil liberties; 2) responsiveness of the government; 3) level of participation; 4) accountability; and 5) rule of law, or rather the validity of the legal system (Barreda 2011: 270).

However, as our research progressed, we modified our initial hypothesis and refocused the study. We...