Investigation of price and capacity decision processes: An empirical study of the United States petrochemical industry

This research combines empirical analysis with survey research to explore decision processes by managers in the oligopolistic petrochemical industry with respect to pricing and capacity adjustment decisions. In order to test the research hypotheses across several businesses, six products were studied. Time series data were obtained for the six products from the first quarter of 1976 through the first quarter of 1991, approximately 65 observations. A survey instrument was distributed to supplement the empirical findings. The survey was sent to 79 key decision makers for the six products and to 26 decision makers for an additional six products as a control group. The usable return rate was over 50 percent.;A positive relationship was found between operating rate and profitability for four of the six products studied. Explanations, based on industry structure, are offered for the two exceptions. No relationship was found between capacity decisions and profitability for five of the six products. However, a relationship was found between profitability and entry/exit decisions in four of the products. Survey research suggests that capacity decisions are driven by expectations of positive net present value. It was found that managers would continue operation during periods of low profitability when the present value of future expected revenue exceeds the present value of expected future costs. For five of the six products, it was found that firms do not optimize profit by producing at the point where marginal revenue equals marginal cost, but at a point above marginal cost. An explanation, based on industry structure, is offered for the exception. Econometric models were developed to predict prices and profitability for the six products studied.;Given the degree of interdependence within the petrochemical industry, it was found that decision processes across different products tend to converge toward common goals. Of 228 tests run on the six products against 38 survey questions, only eight significant differences were found. This is less than would be expected due to random variation alone, providing strong evidence for convergence of the decision processes.