February 2000 - Some say that democracy is more likely to survive under parliamentary governments. That result is not robust to the use of different variables from the Database of Political Institutions, a large new cross-country database that may illuminate many other issues affecting and affected by political institutions. This paper introduces a large new cross-country database on political institutions: the Database on Political Institutions (DPI). Beck, Clarke, Groff, Keefer, and Walsh summarize key variables (many of them new), compare this data set with others, and explore the range of issues for which the data should prove invaluable. Among the novel variables they introduce: · Several measures of tenure, stability, and checks and balances. · Identification of parties with the government coalition or the opposition. · Fragmentation of opposition and government parties in legislatures. The authors illustrate the application of DPI variables to several problems in political economy. Stepan and Skach, for example, find that democracy is more likely to survive under parliamentary governments than presidential systems. But this result is not robust to the use of different variables from the DPI, which raises puzzles for future research. Similarly, Roubini and Sachs find that divided governments in the OECD run higher budget deficits after fiscal shocks. Replication of their work using DPI indicators of divided government indicates otherwise, again suggesting issues for future research. Among questions in political science and economics that this database may illuminate: the determinants of democratic consolidation, the political conditions for economic reform, the political and institutional roots of corruption, and the elements of appropriate and institutionally sensitive design of economic policy. This paper - a product of Regulation and Competition Policy, Development Research Group - is part of a larger effort in the group to understand the institutional bases of poverty alleviation and economic reform. The study was funded by the Bank's Research Support Budget under the research project Database on Institutions for Government Decisionmaking (RPO 682-79). The authors may be contacted at tbeck@worldbank.org, gclarke@worldbank.org, pkeefer@worldbank.org, or pwalsh@worldbank.org
