Differential Efficiency, Market Structure, and Price

A persistent question in industrial economics is the underpinning of the link between market concentration and price. How much of the link can be attributed to market power and how much to market efficiency? This paper develops a theoretical model to address that question. Applied to the US portland cement industry, the model indicates that both impacts matter. In relative terms, however, the market power effect is twice as large as the efficiency effect. An implication for merger policy is that the beneficial efficiency effects of mergers may not be obtained without the detrimental market power effects as well. 1351 1 See for example Ravenscraft (1983). 1352 A. A z z A m & D. R o s e n b A u m i n A p p l i e d e c o n o m i c s 33 (2001) In this paper the link from concentration to collusion and efficiency, and the link from efficiency to concentration are formalized. In so doing, a way to testing Williamson’s (1968) policy tradeoffs, and Demsetz’s efficiency hypothesis is provided. The framework is a hybrid of Azzam’s procedure for separating the market power effect from the cost-efficiency effect of industrial concentration, and Clarke and Davies’ (1982) work on the joint determination of structure and performance. The paper is in six sections. The next section briefly describes the portland cement industry which is used for empirical analysis. The theoretical model is developed in the third section. The model explicitly identifies the concentration-related market power and market efficiency components of price. It also creates a link between firm differential efficiency and market concentration. Section IV describes the empirical approximation of the theoretical model and discusses the data. Results are presented in Section V. These results unambiguously show that rising market power raises price while rising efficiency lowers price. The market power effect, however, is twice the magnitude of the efficiency effect. Hence, overall, increases in concentration increase price. The results also show that concentration is an increasing function of the variance in costs across firms in the market. The greater the cost variance, the more larger firms benefit at the expense of smaller firms and the higher the market concentration. II. The Portland Cement Industry Portland cement is a homogeneous producer good. Given its low ratio of value to weight, it is shipped limited distances.2 Consequently, the US supply of portland cement can be divided into a number of regional markets. Regional markets tend to have a limited number of suppliers. Buying patterns tend to exhibit frequent sales. Many firms are interconnected across regional markets. Previous studies have found evidence of behaviors consistent with tacit collusion in this industry.3 The portland cement production technology uses calcium carbonate rock which is crushed and then combined with lime and sand for grinding. Once ground, the raw materials are fed into a kiln where intense heat causes chemical changes in the composition of the feed stock. Kiln output, clinker, is then combined with gypsum and ground again. The end result is portland cement. There are essentially five variable inputs into cement production: labor; fuel (predominantly coal or natural gas) which is used to heat the kiln and is the primary input expense; electricity, used to operate related auxiliary equipment; feed stocks; and maintenance. Furthermore, as variable inputs are not substitutable, the process exhibits a fixed factor production function. Marginal costs do vary across kilns, however, based on respective technologies, capacities and ages. Portland cement is predominantly used in construction. Over the period 1978–1982, approximately 65% of sales were for building construction, 31% for highway construction, and less than 5% went towards nonconstruction activities.4 The demand for cement should be fairly inelastic since there are few substitutes and its cost makes up a moderate part of most construction projects. Note that asphalt is an alternative for some highway construction applications.