Methodological issues in the projection of households and families in Canada

Abstract

This paper presents some of the methodological and conceptual issues in developing household and family projections. Alternative methods are briefly examined in light of some basic criteria and of Statistics Canada's experience in the application of the headship rate approach. This approach satisfies most of the basic criteria of a robust and practical method, and has the advantage of being able to use the latest population projections by age and sex as its base. It thus takes into consideration the changes in population structure which have a major effect on the number and composition of households and families.

If the projections are prepared for a period of about 20 years, fertility assumptions will have no effect on the projected number of heads. The thrust of the projection work can therefore focus on the extrapolation of the series of headship rates. Other issues covered include conceptual changes in data series; selection of the explanatory variables and categories; and development of alternative assumptions.