Are Political Freedoms Converging?

Article excerpt

I. INTRODUCTION

This article examines the Freedom House (2002) index of political rights and civil liberties in 136 countries to test if political freedoms are converging. Recently, Helliwell (1994), Barro (1996; 1999), and Minier (2001) find evidence that the level of income is a significant determinant of political freedom or democracy. In this article we pose a related question: will political freedoms converge? We address this question empirically by testing if political freedoms converge in the framework of the Solow (1956) growth model. An important implication of Solow's (1956) model is that per capita incomes converge in the long-run steady-state. Given heterogeneous savings rates, population growth rates, and technologies, the Solow (1956) model predicts that per capita incomes will converge "conditionally" to country-specific steady-states or "compensating differentials." Convergence implies that economies with relatively low initial incomes will grow faster compared with economies with relatively high initial incomes to catch up. Numerous empirical studies find evidence that national per capita incomes are conditionally converging (see, e.g., Baumol 1986; Barro 1991; Barro and Sala-i-Martin 1992; Mankiw et al. 1992; Evans 1996; Evans and Karras 1996; Scully and Bass 1998; Li and Papell 1999). Given evidence that political freedom is related to income and a preponderance of evidence that incomes are converging, an interesting question to posit is whether political freedoms are converging.

To perform our empirical tests for convergence, we employ three time-series methodologies. First, we use a simple analysis to determine if the coefficient of variation of political freedom is decreasing over time. Friedman (1992) notes that Hotelling (1933) referred to the coefficient of variation test as "the real test of a tendency of convergence." Next, we follow the approach of Carlino and Mills (1993) and test for stochastic convergence. Stochastic convergence implies that the log of political freedom in country i relative to the world average is a stationary (or trend stationary) process. Given convergence, shocks to freedom in country i relative to the world will be temporary. Structural breaks are allowed in our tests for stochastic convergence. Finally, we utilize a time-series test of [beta]-convergence to determine if estimates of intercepts and trends are consistent with convergence. Convergence implies that countries with freedom levels below the world average will have positive growth in freedom and vice versa.

In spite of some evidence of convergence, freedom levels are not equal for many countries. As a result, we want to analyze factors that might significantly impact the level of freedom. In particular, we examine the relationship between political freedom and the legal system, educational attainment, natural resource dependency, and economic freedom, among other variables, in a cross-sectional analysis. Overall, our findings suggest that political freedoms are converging for approximately half of the countries. Additionally, the legal system, education, economic freedom, and natural resources are significantly related to the level of freedom.

The article proceeds as follows. Section II provides a brief background discussion. Section III describes the data. Section IV discusses our testing methodologies. Section V examines our empirical findings. Section VI summarizes and concludes.

II. BACKGROUND

Hayek (1960) defines freedom as the absence of coercion by government. Political freedom is generally measured in terms of security of political rights and civil liberties. One of the most widely cited indices of political freedom among nations is the annual index of political rights and civil liberties published by Freedom House (2002). These measures of political freedom have been receiving increasing attention in recent years (see, e.g., Barro 1999; La Porta et al. …