|
Article Excerpt
I. INTRODUCTION
In this article, we empirically examine the direction of Granger causality between democracy and rent seeking. In fact, it has been claimed both that democratic regimes may reduce rent-seeking behavior in societies and that rent-seeking behavior may contribute to the downfall of democratic regimes. Since theoretical research shows that given reasonable conditions both are likely to occur, (1) the actual relationship between these two variables has to be uncovered by analyzing it from empirical and historical perspectives. In fact, as Olson (1982) argues, historically democratic political regimes have provided stability and have helped reduce uncertainty in the population by helping to defend societies against roving bandits. This can result in an improvement and a stabilization of the institutional setting of the society by providing clear formal and informal rules and regulations regarding the interaction of actors. Consequently, it has been claimed that rent-seeking behavior will also decrease.
On the other hand, pervasive rent seeking may translate into an increased degree of uncertainty among the population, which may send mixed signals to the market, affecting the productive process, economic performance, the organization of the country, and, ultimately, the democratic process. In fact, a recent wave of empirical research focuses mostly exclusively on the link between institutional quality and economic performance. (2) This is particularly relevant in Latin America, where a clear understanding of the relationship between institutions, economic performance, and political regimes may help explain the seemingly puzzling choice of inward-looking policies by rulers and especially the persistence of such policies even when it was clear that they were doomed to fail. (3) Perhaps the choice of government economic policies and performance has to do with Latin America's particular political setting. According to Taylor, interest groups might have played a role, especially since the politics in the region typically have been marked by dramatic cleavages between different classes and groups. Only when the costs of maintaining inward-looking policies became too high, did such phases end. These factors may help explain both the choice and the persistence of inward-looking development policies in the region, as well as the role played by autocratic governments in the region. (4)
Lack of data has not allowed for an adequate study of the direction of causality between political regimes and institutional quality. The few available studies are either mainly of a descriptive nature or use indirect or subjective data that, despite being widely employed, are controversial. They also tend to use cross-country approaches that may suffer from an omitted variables problem that cannot be controlled by simply applying fixed effects. (5) In fact, Bertrand and Mullainathan (2001) claim that there are several problems with the use of subjective data in economic research. For instance, Mauro (1995), in the specific case of institutional data, argues that evaluators may be influenced by a country's economic conditions as they may assume that a specific institutional aspect cannot be severe if the country is doing well economically. Furthermore, as there are no guidelines as to what constitutes an "expert" evaluator, assessments may be rendered essentially worthless. (6) Therefore, in order to analyze this possible double causality, we use a historical time series of objective rent-seeking data based on discretionary foreign trade regulations that permits us to explore the time series dimension of the link between rent seeking and democracy. (7) We chose the particular case of Uruguay, a small, historically stable, and ethnically and culturally homogeneous Latin American country. As Rama (1991) poses, Uruguay provides an excellent case from the point of view of a rent-seeking society as it represents an extreme case of discretionary trade policies that lasted for decades, first, via an import substitution strategy that was characterized by high protective trade barriers, multiple exchange rates, and an explicit policy that allocated discretionary foreign exchange, approved import licenses, and banned imports that competed with domestic production and, second, via an export promotion strategy which, while not as explicit as the import substitution approach, was nonetheless actively pursued for considerable periods of time.
While we find evidence that democratic processes and rent seeking are Granger-causally linked in both directions, we find that democracies per se are not conducive to improved institutions; democratic longevity, however, is. In fact, not only do we see that the longer the duration of democracy, the lower the discretionary rent-seeking regulations but we also see that the less pervasive the rent seeking in a society. On the other hand, the Granger-causal impact of rent-seeking actions on democracies appears to be quite dramatic regardless of whether the variable of interest is democracy or duration of democracy.
This article is organized as follows. The next section, Section II, describes the panel-autoregressive vector autoregression (VAR) empirical methodology employed in order to assess Granger causality between our variables of interest. Section III describes the data employed and provides basic statistics, and Section IV presents the evidence. Finally, Section V summarizes and concludes.
II. EMPIRICAL METHODOLOGY
In this article, we focus on the dynamic relationship between our variables of interest as well as on the direction of Granger causality and their implied contribution to the possible correlation among these variables. The first step is to analyze the dynamic relationship between rent seeking and democracy. The objective is to examine how the behavior of one given variable is related to the future behavior of the other. This future behavior has two aspects: effect and predictability. The first deals with whether changes in one variable have a lasting impact on another variable. The second examines whether the behavior of one given variable helps predict the behavior of the other. Our methodology consists of estimating and testing VARs in a panel setting that has the following form:...
|
