Econometric modelling of Canadian long distance calling: A comparison of aggregate time series versus point-to-point panel data approaches

Aggregate time series and point-to-point panel data approaches to modelling Canadian long distance calling are compared. It is found that, while both models yield similar sets of aggregate elasticities, the point-to-point panel data models provide more detailed information on consumer behavior. Furthermore, the panel data models take advantage of the disaggregate data available and avoid the problems inherent in aggregate models when the rates of changes in the micro data are different.