Trade Openness and Democratic Survival: An Empirical Investigation.

Attributed to conventional wisdom is the belief that higher levels of international trade promote political stability and democracy. The rationale is that economic openness enhances democratic performance, increasing the probability of regime survival. Many have made such assertions without much theoretical or empirical justification. While policy-makers proclaim the wonders of free trade in promoting democratic values, scholars have addressed the matter in limited ways. This paper attempts to remedy this lack of analytical attention by exploring the relationship between exposure to international trade and the prospects of democratic survival. Drawing on the evidence provided by a binary time series-cross-section (BTSCS) data set of 135 countries during the period 1950-90 (the Przeworski et al 2000 data set), it assesses the conditions under which democracies are more likely to survive. My main finding is that trade openness negatively affects the prospects for democracy. Basically, this paper calls attention to a rich field of research involving the intersection of the literatures on democratization and globalization. Studies that do try to bring the two bodies of literature together suffer from two main shortcomings. First, they focus primarily on the effects of financial liberalization on democracy, rather than on the impact of trade. Second, among the studies that are concerned with trade, the large majority reverses the direction of causality. Instead of examining the influence of trade on democracy, they look into democratic regimes to find explanations for certain levels of trade openness. The model here specified tries to capture the impact of both domestic and international level variables on the process of democratization. It is an improvement over earlier studies in the sense that it tries to control for the international political environment. In addition, this study is methodologically innovative because it imports from the sub-field of International Relations the logit model with temporal dummies that account for temporal independence and applies it in the study of the determinants of democratic survival. The nature of the data poses challenges to the simple logit analysis. BTSCS observations are likely to violate the assumption of temporal independence of ordinary logit models. In this case, the coefficients are biased, and the inferences drawn from an ordinary logit analysis tend to underestimate the standard errors, yielding overly optimistic results. Based on a solution to this problem proposed by Beck, Katz, and Tucker’s “Taking Time Seriously: Time-Series Cross-Section Analysis with a Binary Dependent Variable” (1998), we then estimate a logit model corrected for temporal dependence. Three main advantages of this type of model can be highlighted. First, the proposed technique does not require scholars to learn a totally new methodology to deal with BTSCS data. Second, as Beck, Katz, and Tucker (1998) well point out, a logistic regression corrected for temporal dependence can be readily “combined with Huber (1967) standard errors, which solve other problems inherent in BTSCS data” (p. 1283). Lastly, this method forces researchers to think about the problems of event-history analyses that are not addressed in logistic regressions. [ABSTRACT FROM AUTHOR]