The issue of what determines the level of central government wage expenditures is important in the design of effective reform programs. This paper provides estimates of the impact on central government wage expenditures of (1) institutional factors that influence the level of decentralization of government; (2) the availability of resources; (3) Fund-supported programs; and (A) preference-related variables. These estimates provide a basis for comparing central government wage expenditures across countries in the tradition of Heller and Tait (1984) and possibly for identifying government productive inefficiency or an ambitious set of central government objectives.
Empirical investigation using cross-country time series is carried out based on a simple model of public choice, which suggests a two-equation recursive system consisting of a “total expenditure” and a “wage share” equation. The estimation procedure takes account of the possible bias resulting from correlation between the errors of the total expenditure and the wage share equations. The data span 1980-90 and cover 99 countries.
The paper has five main findings. First, federations and countries with large populations—which are viewed as proxies for fiscal decentralization--have lower central government wage expenditures as a percent of GDP. In addition, central government wage expenditures appear to decline with per capita income, presumably because decentralization and also human capital increase with economic development. Second, general government wage expenditures appear to Increase with per capita income, as indicated by a comparison of industrial countries and developing countries for which general government data are available. Third, the paper finds that private nonguaranteed foreign financing is associated with higher central government expenditures and, indirectly, with higher wage expenditures. Public and publicly guaranteed net financing, provided often for public sector capital projects, are associated with higher total government spending but have only a marginally significant association with wage expenditures. Fourth, medium-term structural adjustment programs, which often include civil service reform measures, have a significant negative association with wage expenditures, while short-term stabilization programs, which do not include such measures, do not. Fifth, there appear to be systematic differences in expenditures on wages across forms of government. Wage expenditures are lower in socialist countries than in democracies, probably because government employees in the former countries receive a large portion of compensation as in-kind benefits. These findings can be used to put countries with diverse characteristics on a comparable basis. They do not necessarily carry policy implications and do not lead to the conclusion that certain forms of government are preferable from the point of view of fiscal consolidation.