A Note on the Statistical Estimation of Supply and Demand Relations from Time Series

IN A discussion in this Journal of the estimation from time series of economic relations such as demand and supply equations, G. Tintner developed a technique based on multivariate methods of statistical analysis, the economic relations being identified with the linear relations existing (within the limits of observational error) among the economic variables considered. In the present note a more precise analysis of the same type is adopted in which some of the limitations of Tintner's procedure are avoided. Some data on demand and supply relations for cotton yarn, recently collected and discussed elsewhere by K. S. Lomax, are reanalyzed to illustrate the proposed method. This example is also used to show how the accuracy of the elasticities determined from the equations may be assessed by the construction of confidence regions>' within which the true values are likely to lie.