Determinants of municipal government expenditures in Alberta using multitude regression analysis
Date
1995
Authors
Rizzo, Rosalia Diana
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
This thesis analyzes the influence of selected expenditure determinants on the supply and demand for police, fire, and recreational services provided by Alberta municipalities. In addition, aggregate expenditures which are the sum of police, fire, and recreational services are analyzed.
The model tested is fitted in semi-log form, where the dependent variable is in logarithmic form and the independent variables are linear, using the method of ordinary least squares regression (OLS). The results are first analyzed using pooled urban and rural data, and secondly separating urban and rural data.
The pooled urban and rural results show that police expenditure tends to be influenced by population characteristics, while fire expenditure tends to be influenced by geographical characteristics. Recreational and aggregate expenditure do not show any specific tendencies. The most important finding in this exercise is the significance of the urban dummy variable in police, fire, recreational and aggregate regressions. This finding shows that urban and rural municipalities differ and should be analyzed separately.
The urban results show distinct differences between cities on the one hand and towns and villages on the other. However, there is very little distinction between towns and villages. This may be due to the population size of cities when compared to towns and villages. The rural results also show a distinction between counties and municipal districts. Other than certain dummy variables, the rural results have only two significant independent variables TG (total grants) and ILI (incidence of low income) in two of the four regressions. The limited number of observations available in the rural analysis may explain the fewer significant independent variables.
Commonalities of the pooled urban and rural results and separate urban and rural results include: a preference for the log-linear functional form and OLS estimation; the unsatisfactory results of the aggregate regression; the insignificance of equalized assessment (defined as the value of property) as an independent variable; and the significance of grants from the province to municipalities.