The estimation of a production function for the construction industry in Malta (1964-1974)

Abstract

There are two alternatives in terms of which production processes can be studied, these being either in terms of cost functions or in terms of production functions. This paper aims at the latter objective, that is the estimation of a production function for the IVial te se Construction industry. As will be explained in Chapter V more than one definition for variables used (especially labour input) were applied and alternative ways adapted to see: (1) which equation would give us positive results (2) which function provides the best fit. Other important tests never carried out on the available data for the period range in question to analyse: - the overall significance of the estimated regression equations - returns to scale - the reliability or non-reliability of both elasticities (that of labour and that of capital) with respect to output. In the case of the C.E.S. in its simplest form an analysis of the Elasticity of Substitution Parameter is given. Like most Cobb-Douglas function studies, parameters are estimated directly from the logarithmically linear production function, thus leaving open an apparent test of the marginal productivity law. In the C.E.S. case we have avoided fitting the non-linear production relation and have instead concentrated on fitting the logarithmically linear marginal productivity equations. Chapter I specifies the economic theoretical background behind an industry's production function, while Chapter II deals with the data used and its interpretation. Appendix to Chapter II gives an analysis of the original results obtained and the reasons why they were discarded. Chapter III gives an account of the general outline on the nature of the Construction industry whilst Chapter V presents a detailed analysis of results obtained. Full account of statistical tests applied in arriving at our conclusions is given in Chapter IV. As can be seen, the sample size was restricted to eleven observations (1964 - 1974). The original purpose was to include more observations, however because of statistical information limitations (see Chapter II) the range had to be shortened.