Assessing the Causes of Capital Account Liberalization: How Measurement Matters


  • Authors’s note: We thank Alexandra Guisinger for encouraging us to pursue this topic. Thanks also to Yong Kyun Kim, James Mahoney, Tom Pepinsky, Steve Nelson, Jason Seawright, Dennis Quinn, and the editor and anonymous reviewers at ISQ for providing extensive comments on earlier drafts of this paper. We also received helpful comments from participants at the Applied Quantitative Methods Workshop at Northwestern University, the 2010 ISA convention, and 2010 MPSA convention. We are grateful to everyone who supplied us with their data sets and helped us replicate their research, particularly Sarah Brooks, Jeff Chwieroth, and Marcus Kurtz. We also thank Menzie Chinn, Hiro Ito, and Ashoka Mody for access to the raw data of their Kaopen index. All remaining shortcomings are, of course, entirely ours. Authors’ names are listed alphabetically. Data and additional materials are available at Steinberg’s Web site,


Karcher, Sebastian and David A. Steinberg. (2012) Assessing the Causes of Capital Account Liberalization: How Measurement Matters. International Studies Quarterly, doi: 10.1111/isqu.12001
© 2012 International Studies Association

Why do countries open their economies to global capital markets? A number of recent articles have found that two types of factors encourage politicians to liberalize their capital accounts: strong macroeconomic fundamentals and political pressure from proponents of open capital markets. However, these conclusions need to be re-evaluated because the most commonly used measure of capital account openness, Chinn and Ito’s (2002) Kaopen index, suffers from systematic measurement error. We modify the Chinn–Ito variable and replicate two studies (Brooks and Kurtz 2007; Chwieroth 2007) to demonstrate that our improved measure overturns some prior findings. Some political variables have stronger effects on capital account policy than previously recognized, while macroeconomic fundamentals are less important than previous research suggests.