Price Leadership and Welfare Losses in U.S. Manufacturing: Comment
评论Gisser的价格领导模型，指出其假设需求弹性为1和领导者与跟随者供给弹性相同存在问题，修改后福利损失可能更大，产出扭曲可能被低估。
Do economists have to worry about the efficiency losses of imperfect competition? Textbooks especially tend to conclude that the answer is negative. The truth is, however, that the relevant literature can be divided in two opposing parts of roughly equal size.' That is clearly an unsatisfactory state and novel contributions employing improved methods are therefore welcome. This decade has witnessed an enormous upswing in oligopoly theory. The deadweight loss literature has been influenced by that to a minor extent only. Avinash Dixit's and Nicholas Stern's (1982) algorithms are however notable exceptions. Another valuable exception is Micha Gisser's (1986) attempt to apply a price leadership model for predicting the order of magnitude of the loss of consumer surplus in the United States. Our intention is, however, to show that a couple of quite reasonable modifications challenge the conclusion of small deadweight losses implied by the latter approach. We do not suggest any improvement allowing for the influence of product differentiation and concentration of local markets; we only note that the deadweight losses are under rather than overestimated by those omissions. Our contribution is to show that two other assumptions that appear in Gisser's analysis are crucial for the results. We find that the ideas of unitary demand elasticity and equal supply elasticity for the leaders and the price-taking fringe are questionable. We also point out that the output distortion (percentage loss of output) may be quite large even if the percentage loss of consumer surplus is small. This fact seems to be ignored in the literature. Consider first demand elasticity. Gisser assumes it to be unity, like Arnold C. Harberger (1954). Empirical estimates are not necessarily reliable, but the fact that they are often below unity should in our opinion be taken seriously. (See, for example, M. S. Houthakker and Lester P. Taylor, 1966, and for the U.S. food and tobacco industries, Emilio Pagoulatos and Robert Sorensen, 1986). Gisser's technique of predicting welfare losses means that realistic values of demand elasticity and concentration are inserted in a formula that is derived from a theoretical model. It is therefore inadequate to limit its applicability to the case of unitary elasticity, which appears dubious from an empirical standpoint. Consider then supply elasticity. Gisser assumes that both the leading group and the fringe of price-taking firms have the same supply elasticity. Our intention is not to discuss the behavior of firms in detail, but we nevertheless find it questionable that firms assume the roles as either active leaders or passive price takers, if their technologies are more or less identical. Why do firms not dominate because they have a larger supply elasticity? It could even be considered a styl*Acting professor, Department of Economics, Abo Akademi (The Swedish University of Turku), Fanriksgatan 3, SF-20500 Turku, Finland. Participants in the Department's research seminars have delivered useful criticism. I thank Leila Stahl for research assistance. Remaining errors are my own. 'Losses that are less than 1 percent of GNP are reported by Arnold C. Harberger, 1954, David Schwartzman, 1960, Frederick W. Bell (1968), Dean Worcester, 1973, John Siegfried and Thomas K. Tiemann, 1974, Bjmrn Wahlroos, 1984, and outside the United States, Wahlroos (1984; Finland). Losses that are larger than 1 percent and usually larger than 3 percent are reported by David Kamerschen, 1966, and Keith Cowling and Dennis C. Mueller, 1978, for the United States. The losses may be large in the UK according to Cowling and Mueller, 1978, and Malcolm C. Sawyer, 1980, and in France according to Frederick Jenny and Andre-Paul Weber, 1983. Simulations in different types of models by Abram Bergson, 1973, J. A. Kay, 1983, and G. K. Yarrow, 1986 suggest losses that are at least potentially large.