The specifications and testing of structure-conduct-performance relationships in Australian manufacturing

Beginning with Mason (1939) and then continuing perforce with Bain (1956 and 1968) into the early 1970s, research in the mainstream of industrial organisation (10) was dominated by the structure-conduct-performance paradigm. The paradigm relates an industry's structural characteristics to the conduct patterns of its constituent firms, which determines an industry's performance. Specification of the relationships in this area is often largely ad hoc and its testing is subject to a number of difficulties. Considering an equilibrium oligopoly framework, mainly based on Cowling and Waterson (1976) study, the thesis derives short-run equilibrium profit, concentration and industry conjectures models. The empirical analysis will consist of two parts. Firstly, the above three equilibrium models are tested with different versions both in linear and logarithmic functional forms. Empirical testings includes testing the major criticisms against structure-conduct-performance studies with the profit model, distinguishing consumer-producer goods in explaining the structure-performance relationships and comparison of the hypothesised in4dels with ad hoc versions of profits and concentration. The profit model includes concentration, elasticity of demand, conjectures, product differentiation dummy, import intensity and capital intensity as explanatory variables. Concentration is determined by number of firms, elasticity of demand, product differentiation dummy, cost disadvantage ratio of small firms and conjectures. Industry conjectures is determined by levels and stability of concentration, growth in sales, product differentiation and import intensity. For each model, both linear and logarithmic functional forms (with some intermediate stages) are tested for empirical purposes. For the equilibrium profit model, oligopolistic structural and behavioural variables, viz, concentration, conjecturess, elasticity of demand and capital intensity are found to be significant in most of the cases. For the equilibrium concentration model, number of firms, elasticity of demand and cost disadvantage ratio of small firms are found to be significant in most cases. For the equilibrium conjectures model, none of the above factors, viz. levels and stability of concentration, growth in sales, product differentiation and import intensity is found to be significant in explaining conjectures. The empirical analysis deals with testing the importance of major criticisms against the structure-conduct-performance studies in explaining the relationships. These criticisms are: i) omission of the relevant explanatory variables, ii) measurement error in the variables, and iii) aggregation bias and iv) existence of simultaneous causality among variables and v) mis-specification of the functional form of the model. These are tested with the profit models. The first three criticisms, viz, errors of omitting relevant variables, measurement error and aggregation bias are tested with the profit model. Resulting estimates suggest the importance of dealing with each of these issues in explaining the structure-performance relationship. The existence of simultaneous relationships among the variables is tested with two systems of equations. The first system of equations consists of equilibrium profit, concentration and conjectures models. The second system of equations consists of disequilibrium profit and concentration models and the equilibrium conjectures model. Profits, concentration and conjecturess among firms are treated as endogenous variables. For both systems, the empirical findings support there is a simultaneous relationship between the profit and concentration equation. The second system of equations deals with the existence of sub-optimal behaviour. This is done by specifying disequilibrium models separately for profit and concentration. Partial adjustment model is considered with a constant rate of adjustment towards the equilibrium level, assuming that any deviation from equilibrium is corrected at the constant rate of adjustment. For the profit model, the signs of coefficients of the disequilibrium version are found to be same as the equilibrium model with a significant positive sign for initial profit. Also for the concentration model, the signs of the coefficients of the disequilibrium version are found to be same as the equilibrium model with a significant positive sign for initial concentration. The annual adjustment rate of profit and concentration are both found to be slow. For the equilibrium profit, concentration and conjectures models, empirical analysis is carried out against a sample of 102 Australian manufacturing industries at the Australian Standard Industrial Classification (ASIC) at the four-digit level data for 1984/85. For the disequilibrium profit and concentration model, the data are used for two periods (1977/78 and 1984/85), with a seven year lag for the above industries.