THIS is a reply to an article of M. A. Adelman, The Measurement of Industrial Concentration, which appeared in the November I95I issue of this REVIEW. Comments by Edwards, Stocking, George, and Berle appeared in the May I952 issue. This reply will endeavor to avoid duplicating points already made by the other discussants, and therefore should be read in conjunction with their contributions in order to provide a full critique of the Adelman article. Adelman's article is broken down essentially into three sections: Parts I and II discuss conceptual problems of measurement; Part III is concerned with concentration during and immediately following World War II; and Parts IV-VI examine the long-term trend of concentration. This reply will take up each of these three sections in order.

P THE major part of Dr. Backman's criticism of my article, Means, Thorp, and Neal on Price Inflexibility, is concerned with my analysis of Thorp. The essence of my position was that Thorp's data, by their very nature, suffered from three and possibly four biases; that these biases operate against appearance of any relationship between concentration and price rigidity; and that they do not appear to be offset by any biases operating in opposite direction. Dr. Backman offers a new body of data which concerns bias resulting from Thorp's inclusion of products . . such as meat, whose costs are made up largely of raw materials and whose prices may therefore be expected to fluctuate more or less in accordance with changes in prices of their raw materials. Inasmuch as prices of these products tend to be highly flexible, question at issue is their level of concentration. If they tend to have high concentration ratios, result is to provide support to my position that their inclusion in Thorp's study tends to operate against appearance of a relationship between concentration and price rigidity. Using information compiled by Means on industries and applying it to Thorp's products, Backman finds that 'raw-materialaffected' products tend to have high concentration ratios. Of 9 I products, I9 had a concentration ratio of over 8o per cent, and 42 products had a ratio between 50 per cent and 8o per cent -or a total of about two-thirds of products in this category. Dr. Backman devotes a considerable amount of attention to bias resulting from Thorp's inclusion of products which are on a delivered price basis in nation as a whole, or at least a broad geographic area and have high ratios of freight costs to delivered prices. Presumably, during depression producers in such industries tended to increase proportion of their shipments made to distant areas, thereby reducing their realized prices. As to incidence of bias within Thorp's of a cross-section, I had listed as examples iron and steel, chemicals, and non-ferrous metals (all highly concentrated products). I added that it is difficult to find within sample any products of this type which have low concentration ratios. Backman's criticism does not go to incidence of bias, but only to its extent within iron and steel, chemicals, and non-ferrous metals. This is done through comparisons of Census price declines with BLS price declines. Although in two of three groups (iron and steel and chemicals) ' he finds Census price declines to have been greater than those of corresponding BLS series (thus again buttressing my position), not too much weight should be given to these comparisons because of differences in product definitions. The next bias stems from understatement of concentration in market resulting from use of concentration ratios on a national basis for products sold largely on a local or regional basis. Here, Backman conveys impression that class of products which I had cited (stone, clay, and glass) represents totality of products affected by bias. His conclusion, of course, is that they are so few in number as not to alter the general conclusion of Crowder-Thorp study. This, despite fact that I had specifically labelled these products as merely illustrations of bias.2 With respect to fourth bias, i.e. shift during a downswing to lower-priced items within a Census product (which can occur because of excessive breadth of many of Census product definitions), Backman quotes and then attacks an hypothesis which I had advanced to effect that bias may be more important in products of high than of low concentration. In quoting me, however, he omitted sentences setting forth central reason underlying hypothesis; effect is to reduce my argument to level of assertion. The full passage quoted

DURING the late I930'S the between economic concentration and depression price behavior was a matter of considerable controversy. At the heart of the controversy were three elaborate statistical works.' The first, by Gardiner Means, found that a rough relationship existed between economic concentration and price rigidity; the second, by Willard Thorp and Walter Crowder, concluded that there was no such and that indeed price rigidity could be explained by the characteristics of the product (durability, use, etc.); and the third, by Alfred Neal, found that differences in price change are associated with differences in changes of direct costs and are therefore not to be attributed to differences in concentration. This article represents an attempt at a critical evaluation of these studies.

John M. Blair, Harrison F. Houghton, The Lintner-Butters Analysis of the Effect of Mergers on Industrial Concentration, 1940-1947: A Reply, The Review of Economics and Statistics, Vol. 33, No. 1 (Feb., 1951), pp. 63-67