An empirical analysis of the determinants of capital and current account restrictions and their liberalization

The primary objective of this thesis is to empirically explore the characteristics associated with the policy choice to impose capital and current account controls. Its approach is to analyze a large group of countries over time so as to uncover the underlying determinants of capital and current account restrictions. It is well known that both capital and current account controls come at a cost. Why then, are they so frequently used? Are they justified on purely economical grounds, or do political motivations play an important role? Do institutional factors such as central bank independence matter for these controls? We pursue the question: is it possible to identify the common political, economic and institutional characteristics of the countries that adopt these controls and the time periods in which they are adopted? These issues are analyzed in the three empirical chapters which provide the main substance of the thesis.; Before we proceed with the empirical pieces, we begin with an introductory chapter. Chapter two is a descriptive chapter on the types and dynamics of controls. Finally, chapter three provides a short literature review on capital controls.; The first paper (chapter four) provides a simple analysis of determinants of capital account controls. That paper mainly expanded on previous papers in the literature that dealt with the same issue. In the second paper (chapter five), we analyze the determinants of current account restrictions in addition to the capital account restrictions. We also use a more detailed formulation of the independent variable in the regressions than the one we used in the first paper. Finally, the third empirical piece (chapter six) is a short note that attempts to shed some light on the factors that play a role in external sector liberalization. In particular, we explore whether "peer behavior" has an impact on a country's decision to liberalize.; In the regression analysis, we use panel data and our dependent variable is qualitative. The regression method is maximum likelihood logistic estimation.