On the Problem of Missing Measurements in the Estimation of Economic Relationships
Jan Kmenta · Deep Blue (University of Michigan) · 1978
This paper deals with the problem of estimating the parameters of economic relationships when some of the observations are incomplete but the sample selection rules are not violated. The models considered include classical multiple regression, generalized regression (with special attention to autoregressive disturbances and seemingly unrelated regressions), and recursive systems. In each case we examine the possibility of extracting information from observations with missing measurements and then analyze suitable estimation methods.