Do Constititutions Cause Large Governmets?

Quasi-Experimental Evidence

by

Torsten Persson and Guido Tabellini

Abstract

How do constitutional rules for elections and legislation affect the size of government? We ask this question in a new sample of about 80 countries in the 1990s. In addition to conventional regression methods, we use quasi-experimental, matching methods, which more convincingly address legitimate criticisms of causal inference from cross-country data. Both sets of estimates suggest that presidential regimes and majoritarian elections produce smaller governments.