ABSTRACT
This situation is particularly regrettable because the empirical validity of the models arising out of public choice assumptions is far from evident. Even the very core of the theory - the proposition that government spending would be higher where benefits are concentrated and costs are diffuse - fares badly in a simple confrontation with cross-national data. For example, the public choice theory implies [Fiorina and Noll, 1978: pp. 252-253] that ceteris paribus governments should have grown larger in the plurality/majority than in the proportional representation systems. A quick calculation shows that between roughly 1960 and 1979 total government expenditures increased on the average by 8.5 per cent in five plurality/majority systems and by 18.3 per cent in ten proportional representation systems; by 1977-79 the average level of total government expenditures was 38.8 per cent of GDP in the first group of countries and 49.1 in the second. [Expenditure data are from Schott, 1984: Table 3.6; information about electoral systems is derived from Rae, 1971: Table 2.1.] Moreover, government expenditure turns out to be higher by seven per cent in countries that have a unitary rather than a federal system [Saunders and Klau, 1985: p.117]. When one reads the American literature on government "overspending" one tends to forget that during the sixties and seventies total government expenditures grew less in the United States than in any other industrialized country and that among the OEeD countries only Japan and Australia now have lower government expenditures.
