Testing Multiple Non-nested Factor Demand Systems

Empirical factor demand analysis typically involves making a choice from among several competing non-nested functional forms. Each of the commonly used factor demand systems, such as Translog, Generalized Leontief, Quadratic, and Generalized McFadden, can provide a valid and useful empirical description of the underlying production structure of the firm. As there is no theoretical guidance on selecting the model which is best able to capture the relevant features of the data, formal testing procedures can provide additional information. Paired and joint univariate nonnested tests of a null model against both single and multiple alternatives have been discussed at length in the literature, whereas virtually no attention has been paid to either paired or joint multivariate non-nested tests. This paper shows how some multivariate non-nested tests can be derived from their univariate counterparts, and examines how to use these tests empirically to compare alternative factor demand systems. The empirical application involves the classical Berndt- Khaled annual data set for the U.S. manufacturing sector over the period 1947-1971. A statistically adequate empirical specification is determined for each competing factor demand system. The empirical results are interpreted for each system, and the models are compared on the basis of multivariate paired and joint non-nested procedures.