Choice of Inputs and Outputs for Production Analysis

Production creates value by transforming inputs into outputs. Classification of variables as inputs or outputs depends on the scope of decision making by the firm. Inputs enter the boundary of the firm from outside without any prior processing by the firm and once transformed into outputs exit its jurisdiction without any further processing. This chapter highlights the defining characteristics of inputs and outputs both for single stage and multi-stage production. The necessary conditions that must be met for a valid aggregation of several inputs into total expenditure are discussed both for nonparametric and parametric models of production. Several statistical tests of hypotheses related to aggregation of several inputs or exclusion of individual inputs in nonparametric models are discussed. The technology set of feasible input-output bundles invariably depends on many environmental or contextual variables that are outside the control of the producer. In parametric Stochastic Frontier Analysis they can be directly included as determinants of the mean or variance the technical efficiency factor causing shifts in the production frontier. In nonparametric Data Envelopment Analysis influence of such factors is measured through a second stage regression of efficiency scores on the contextual variables. The alternative approaches of a second stage least squares regression and a truncated regression are briefly discussed. The chapter ends with examples of input-output choice in several popular areas of application like manufacturing, banking, and health care.