Constitutional safeguards against centralization in federal states: An international cross-section analysis

Abstract

This paper contains an international cross-section analysis of the share of central government expenditure in total government expenditure for a sample of about 50 countries and a subsample of 23 industrial countries in 1989–91. The expenditure shares, their changes and the unexplained residuals for each country are reported in Table 1. As the analysis demonstrates, the share of central government is significantly lower, if income per capita and the country's area are large and if it is a federal state. The explanatory power of the equation rises considerably if the binary dummy for federalism is replaced by quantitative constitutional variables. The most powerful single explanatory variable is the age of the constitutional court in the complete sample or the constitutional court's independence of union institutions in the sample of industrial countries. The equation's explanatory power (adjusted for degrees of freedom) can be raised by allowing also for the degree of control which provincial institutions have over the constitution and over the second chamber and by taking into account whether an increase in federal tax rates requires a popular referendum. Other types of constitutional referenda and the relative age of the federal constitution do not seem to matter. Among the federal states, the share of central government is much larger than predicted in the United States and Mexico, and it is much smaller than predicted in Argentina and Canada. The constitutional variables are particularly helpful in explaining the relatively small share of central government in Switzerland, Malaysia, Germany and Austria. The last section draws conclusions for the design of constitutions with some special applications to the European Union.