The author is Assistant Professor of Economics, University of Michigan. For their advice and counsel, he wishes to thank Arcadius Kalian and Donald McCloskey. Lon Peters, George Stigler, the Economic History Workshops at the Universitites of Chicago and of Michigan, and two anonymous referees also provided helpful comments and criticism. He also thanks several German firms and their archivists, who generously allowed him to pore over their collections: Lutz Hatzfeld (Mannesmann), Bodo Herzog (Gutehoffnungshutte), Renate Kohne (Krupp), and Gertrud Milkereit (Thyssen). The German Academic Exchange Service and the Social Science Research Council provided financial support for this research.

¹ Oberschlesischer Berg- und Huttenmannischer Verein, petition of 10/31/1901, Zentrales Staats-archiv, Merseburg, 120, c VII, 1, 42, 36, Bl. 151; Eisenhuttenleute, Verein Deutscher, ed, Eisen-huttenwesen, 7th ed. (Dusseldorf, 1910), pp. 224–25Google Scholar.

² In both the legal and original theoretical contexts, dumping refers primarily to selling exports below the domestic price, not necessarily below costs; see General Agreement on Tariffs and Trade (Geneva, 1974)Google Scholar, Article IV, and Viner, Jacob, Dumping: A Problem in International Trade (Chicago, p. 3Google Scholar.

³ H. A. Bueck, et al., to members of the Statistisches Verband der Roheisenproduzenten, Dusseldorf, March 4, 1879, Gutehoffnungshiitte Historisches Archiv, GHH 300 0039/2.

⁴ For reasons of data availability the years covered in this calculation are 1883–1890, 1894–96, 1900–02, 1904–13. Gutehoffnungshutte Historisches Archiv, GHH 300 1323/011; 300 108; Statis-tisches Jahrbuchfiir das Deutsche Reich, 15 (1894), p. 128Google Scholar ; Statistik des Deutschen Reichs, Vierteljahre-sheft (1914), pt. I, p. 84.

⁵ Hübner, Erhard, Die deutsche Eisenindustrie - ihre Grundlagen, ihre Organisation und ihre Politik (Leipzig, 1913), pp. 105–20Google Scholar; Krupp Historisches Archiv, WA VI 1172; August-Thyssen-Hütte, ATH-(A/1449).

⁶ Stahl und Eisen regularly published the production quotas and actual sales of the Federation.

⁷ Innern, Reichsamt des, Kontradiktorische Verhandlungen über Deutsche Kartelle, 5 vols. (Berlin, 1904–1905), vol. 3, pp. 14, 431–32Google Scholar, and vol. 4, pp. 257–58.

⁸ Germans already sold substantial amounts of rails in Britain by the mid-1890s and in Belgium after 1900. Apparently only the high tariffs of the other Association members, France and the U.S., shielded them.

⁹ Innern, Reichsamt des, Kontradiktorische Verhandlungen, vol. 4, pp. 286–90Google Scholar.

¹⁰ Neuburger, Hugh, “German Banks and German Economic Growth from Unification to World War I” (Ph.D. diss., Univ. of Chicago, 1974), p. 70Google Scholar ; Riesser, Jacob, The German Great Banks and their Concentration, U.S. Senate, National Monetary Commission Publications, vol. 14 (Washington, D.C., 1911), pp. 725–49Google Scholar.

¹¹ Blaich, Fritz, Kartell- und Monopolpolitik im Kaiserlichen Deutschland: Das Problem der Markt-macht im deutschen Reichstag zwischen 1879 und 1914 (Dusseldorf, 1973), pp. 51–60, 109–26, 152Google Scholar; Hardach, Karl W., Die Bedeutung wirtschaftlicher Faktoren bei der Wiedereinfuhrung der Eisen- und Get-reidezolle in Deutschland 1879 (Berlin, 1967), pp. 38–39, 64Google Scholar; Sonnemann, Rolf, Die Auswirkungen des Schutzzolls auf die Monopolisierung der deutschen Eisen- und Stahlindustrie, 1879–1892 (Berlin, 1961), pp. 26–41Google Scholar; Stolper, Gustav, German Economy, 1870–1940: Issues and Trends (New York, 1940), p. 85Google Scholar.

¹² For instance, Schaltenbrand, chairman of the Steel Mill Federation, stated that the cartels were necessary in order for the tariffs to benefit the German steel industry, as they were intended. Innern, Reichsamt des, Kontradiktorische Verhandlungen, vol. 4, p. 235Google Scholar.

¹³ Liefmann, Robert, Kartelle und Trusts und die Weiterbildung der volkswirtschaftlichen Organ-isation. (Stuttgart, 1910), pp. 22–23Google Scholar; Maschke, Erich, Grundzüge der deutschen Kartellgeschichte bis 1914 (Dortmund, 1964), p. 16Google Scholar; Treue, Wilhelm, Gesellschaft, Wirtschaft und Technik Deutschtands im 19. Jahrhundert (Munich, 1975), pp. 249–52Google Scholar.

¹⁴ Bohme, Helmut, Deutschlands Weg zur Grossmacht: Studien zum Verhdltnis von Wirtschaft und Staat Wahrend der Reichsgrundungszeit, 1848–1881 (Cologne, 1966), pp. 530–86Google Scholar; Epstein, Klaus, “The Socio-Economic History of the Second German Empire,” Review of Politics, 29 (01 1967), 100–12Google Scholar; Eyck, Erich, Bismarck and the German Empire (London, 1958), pp. 258–59Google Scholar; Gerschenkron, Alexender, Bread and Democracy in Germany (Berkeley, 1943), pp. 51–53Google Scholar; Lambi, Ivo N., Free Trade and Protection in Germany, 1868–1879 (Wiesbaden, 1963)Google Scholar; Passant, E. J., A Short History of Germany, 1815–1945 (Cambridge, 1969), pp. 92–93, 120–23, 180–81Google Scholar; Wehler, Hans-Ulrich, Bismarck und der Imperia-lismus (Cologne, 1969), pp. 486–89Google Scholar.

¹⁵ H.A. Bueck, GHH 300 0039/2;, Maschke, Grundzüge der deutschen Kartellgeschichte, p. 16Google Scholar.

¹⁶ Nussbaum, Helga, “Zentralverband Deutscher Industrieller 1879–1919,” in Fricke, Dieter, ed., Die Burgerlichen Parteien in Deutschland: Handbuch der Geschichte der burgerlichen Parteien und an-derer burgerlichen Interessenorganisationen vom Vormdrz bis zum Jahre 1945 (Berlin, 1968), vol. 2: 855–57Google Scholar.

¹⁷ Innern, Reichsamt des, Kontradiktorische Verhandlungen, vol. 3, pp. 196, 245Google Scholar; vol. 4, p. 235; , Blaich, Kartell- und Monopolopolitik, p. 124Google Scholar.

¹⁸ , Stolper, German Economy, p. 37Google Scholar.

¹⁹ Clapham, John H., The Economic Development of France and Germany, 1815–1914, 4th ed. (Cambridge, 1936), pp. 317–22Google Scholar; Sombart, Werner, Der Moderne Kapitalismus, 3 vols. (Munich, 1928), vol. 3, p. 64Google Scholar.

²⁰ Lütge, Friedrich, Deutsche Sozial- und Wirtschaftsgeschichte (Berlin, 1966), p. 519Google Scholar.

²¹ Innern, Reichsamt des, Kontradiktorische Verhandlungen, vol. 3, p. 196Google Scholar.

²² Stahl und Eisen, 23 (1912), 2, 771Google Scholar; Statistisches Jahrbuch, 35 (1914), 112Google Scholar.

²³ See also Corden, W.M., The Theory of Protection (Oxford, 1971)Google Scholar, and Michaely, Michael, The Theory of Commercial Policy (Chicago, 1977)Google Scholar.

²⁴ Leisse, Wilhelm, Wandlungen in der Organisation der Eisenindustrie und des Eisenhandels seit dem Griindungsjahre des Stahlverksverband (Munich, 1912), p. 162Google Scholar; Zollner, August, Eisenindustrie und Stahlwerksverband (Leipzig, 1907), pp. 194–95Google Scholar; Krupp Historisches Archiv, W.A., IV 1055; IV 1172; IV 1597; X a3 177.

²⁵ Three-year averages were used to smooth over short-term fluctuations. The five chosen three-year periods offered the advantage either of proximity to an industrial census or of an abundance of other data. The 1900–01 period was atypical in that the iron and steel cartels, following the lead of the coal cartel, seem to have tested the strength of their monoply power then.

²⁶ From Table 1 and Hoffmann, Walter G., Das Wachstum der deutschen Wirtschaft seit der Mine des 19. Jahrhunderts (Berlin, 1965), pp. 390, 454–55, 598–601Google Scholar.

²⁷ Two of the more extended discussions are in Sonnemann, Die Auswirkungen des Schutzzolls and in Wilhelm Leisse, Wandlungen in der Organisation der Eisenindustrie und des Eisenhandels.

²⁸ Innern, Reichsamt des, Kontradiktorische Verhandlungen, vol. 3, p. 138Google Scholar.

²⁹ Eisen, Verein deutscher and Industrieller, Stahl, Statistik (Düsseldorf, 1887–1902)Google Scholar; Gutehoff-nungshOtte Historisches Archiv, GHH 300 108.

³⁰ Statistisches Jahrbuch, 15 (1894), 24–26Google Scholar; 25(1904), 49–51.

³¹ McCloskey, Donald N., Economic Maturity and Entrepreneurial Decline: British Iron and Steel, 1870–1913, Harvard Economic Series, vol. 142 (Cambridge, Mass., 1973), pp. 35–38Google Scholar; , Blaich, Kartell-und Monopolpolitik, p. 112Google Scholar; , Maschke, Grundzüge der deutschen Kartellgeschichte, p. 29Google Scholar.

³² Schumpeter, Joseph A., Capitalism, Socialism and Democracy (New York, 1942), p. 104Google Scholar.

³³ Krawinkel, Max L., Die Verbandsbildung in der deutschen Drahtindustrie (Cologne, 1968), p. 10Google Scholar; Liefmann, Robert, Kartelle und Trusts und die Weiterbildung der volkswirtschaftlichen Organisation, 2nd ed. (Stuttgart, 1924), p. 47Google Scholar. See also Rips, Franz, Die Stellung der deutschen Eisenindustrie in der Aussenhandelspolitik 1870 bis 1914 (Jena, 1941), pp. 40–46Google Scholar.

³⁴ Macrosty, H. W., The Trust Movement in British Industry (London, 1907)Google Scholar; Carr, James C. and Taplin, W., History of the British Steel Industry (Cambridge, Mass., 1962), pp. 254–62Google Scholar; McCloskey, Economic Maturity, chap. 2.

³⁵ Burn, Duncan, The Economic History of Steelmaking, 1867–1931: A Study in Competition (Cambridge, 1940), p. 213Google Scholar; Burnham, T. H. and Hoskins, G. O., Iron and Steel in Britain, 1870 to 1930 (London, 1943), pp. 38, 45, 187Google Scholar; , Carr and , Taplin, A History of the British Steel Industry, pp. 172–73Google Scholar; Kindleberger, Charles P., “Germany's Overtaking England, 1806–1914,” part 2, Weltwirtschaftliches Archiv, 111 (1975), 477–504Google Scholar ; Landes, David S., The Unbound Prometheus: Technological Change and Industrial Development in Western Europe from 1750 to the Present (Cambridge, 1969), pp. 262–69Google Scholar; Temin, Peter, “The Relative Decline of the British Steel Industry, 1880–1913,” in Rosovsky, Henry, ed., Industrialization in Two Systems (New York, 1966), pp. 140–55Google Scholar.

³⁶ McCloskey argued that British total productivity in the early twentieth century was virtually indistinguishable from American for both pig iron and steel (Economic Maturity, chap. 7). Robert C. Allen's 1977 article found British blast furnaces in 1909 to be 5 percent less efficient than American and 2 percent more efficient than German (“The Peculiar Productivity History of American Blast Furnaces, 1840–1913,” this Journal, 37 [09 1977], 605–33)Google Scholar. Allen's 1979 article finds the Germans and the Americans both 15 percent more efficient at steelmaking than the British in 1907–09 (“International Competition in Iron and Steel, 1850–1913,” this Journal, 39 [12 1979], 911–38)Google Scholar. Peter Berck calculates that American blast furnaces were about 5 percent more efficient than the British in 1890 (“Hard Driving and Efficiency: Iron Production in 1890,” this Journal, 38 [12 1978], 879–900)Google Scholar.

³⁷ Statistik des Deutschen Reichs, Vierteljahresheft (1914), pt. I, pp. 358 ffGoogle Scholar; Gutehoffnungshutte His-torisches Archiv, GHH 213421/5–19; Gutehoffnungshutte, 1810–1910 (Oberhausen, 1910), table AGoogle Scholar.

³⁸ , McCloskey, Economic Maturity, pp. 77–78Google Scholar.

³⁹ Statistisches Jahrbuch, 31 (1910), 98Google Scholar; 35 (1914), 110; 40 (1919), 117; Mineral Statistics, Parliamentary Papers (1914–1916), vol. 40, p. 270Google Scholar; , McCloskey, Economic Maturity, pp. 77–84Google Scholar; , Burnham and , Hoskins, Iron and Steel, p. 315Google Scholar.

⁴⁰ , McCloskey, Economic Maturity, p. 86Google Scholar; Shepherd, Ronald W., Theory of Cost and Production Functions (Princeton, 1970), p. 176Google Scholar.

⁴¹ , McCloskey, Economic Maturity, pp. 120–27Google Scholar.

⁴² Webb, Steven B., “The Economic Effects of Tariff Protection in Imperial Germany, 1879 to 1914” (Ph.D. diss., Univ. of Chicago, 1978), pp. 31–33Google Scholar.

⁴³ The productivity differentials ranged from 6 to 27 percent; Ibid., pp. 32–34. The weakness of this estimate is that trade associations in Britain also distorted prices and that Gutehoffnungshutte may not represent average German productivity.

⁴⁴ Statistik des Deutschen Reichs, Auswartiger Handel, 271 A (1913), pt. XI, p. 56Google Scholar; Vierteljahresheft (1914), pt. II, p. 88.

⁴⁵ The cost of shipping pig iron in large lots from Middlesborough-on-Tees to Antwerp averaged 4.13 marks per ton in 1912–13; Stahl und Eisen, 33 (1912), 75, 629, 1153, 1716Google Scholar; 34 (1913), 79, 618, 1170, 1708. Shipping charges from the Ruhr to Antwerp for steel products in 1904 were 4.50 marks per ton; Gutehoffnungshutte Historisches Archiv, GHH 300 0030/0, Bl. 237. Since the Antwerp to Middlesborough rate rose by 15 percent from 1904 to 1912–13,1 assume that the freight rates on the Rhine did the same and rose to 5.1 marks per ton, for a total of 9.25 marks per ton. Shipping costs for steel would have averaged considerably above this, because water freight for steel goods was about two-thirds greater that for pig iron (Clapp, Edwin J., The Navigable Rhine [New York, 1911], p. 115Google Scholar), and because water freight rates for steel goods from the Ruhr to Antwerp were for the most favorably located plants, not the average ones. The input-output ratios probably make a small upward bias in the estimate of British input costs and, hence, of British productivity, because the Germa n input-output ratios would in general be less efficient than the optimal with British relative input prices.

⁴⁶ , Landes, The Unbound Prometheus, pp. 334, 352–54Google Scholar. See McCloskey, Economy Maturity, chap. 1, for an extensive discussion of these accounts. Peter Temin presents the only explanation for inferior British productivity that does not assume entrepreneurial laxity (“The Relative Decline”). He notes that slower demand growth in traditional British markets would slow their rate of investment. That would make their capital stock older on average, and therefore less efficient.

⁴⁷ Not all innovations involved vertical integration. For instance, the change-over to making tin-plate from steel instead of wrought iron resulted in vertical disintegration. Tinplate was a specialty of the British, however, not of the Germans; , Carr and , Taplin, British Steel Industry, pp. 114–18Google Scholar.

⁴⁸ Oswald Gellert indicates that the capital-output ratio of a Thomas-Bessemer mill was at least 5 times that of an open hearth mill ( Eisen und Alteisen [Munich, 1912], p. 67)Google Scholar.

⁴⁹ A simple numerical example illustrates the possibilities. Suppose the cost function of a large mill is C L = 20 + Q3 and that of a small mill is C s = 1 + 50 Q2. The minimum average costs are 13.9 for the large and 14.14 for the small mill, so the large mill is more efficient in the usual sense. It takes 15.23 small mills operating at the minimum cost level to produce as much as one large mill. If the price alternates between 15 and 13, and if firms produce where marginal cost equals price, the large mill makes money on average and the small mills lose. The large mill makes more in the good years and loses less in the bad years. If the price alternates between 10 and 18, on the other hand, large and small mills both profit on average, but the small mills profit more, chiefly because their losses are smaller in the bad years. If the price is 17 in two out of three years and 10 in the other, both make positive profits on average, and the large mills have higher average profits. But the large mills do suffer larger losses in the bad years, making them the riskier investment.

⁵⁰ lnnern, Reichsamt des, Kontradiktorische Verhandlungen, vol. 3, pp. 17, 19Google Scholar; vol. 4, pp. 448–49 (Kirdorf), 461 (Bueck); , Rips, Die Stellung der deutschen Eisenindustrie, p. 43Google Scholar; , Krawinkel, Die Ver-bandsbildung in der deutschen Drahtindustrie, p. 10Google Scholar.

⁵¹ Krupp Historisches Archiv, WA, VII1947.

⁵² Parliamentary Papers (1914–1916), vol. 80, p. 270Google Scholar.

⁵³ Statistisches Jahrbuch, 15 (1894), 24Google Scholar; 25 (1904), 49; 31 (1910), 98.

⁵⁴ Pounds, Norman G. and Parker, William N., Coal and Steel in Western Europe (London, 1957), pp. 136–39Google Scholar; , Carr and , Taplin, British Steel Industry, pp. 175–77Google Scholar.

⁵⁵ , Allen, “The Peculiar Productivity History,” p. 613Google Scholar; , Carr and , Taplin, British Steel Industry, pp 244–45Google Scholar.

⁵⁶ , Clapham, Economic Development of France and Germany, p. 238Google Scholar; , Carr and , Taplin, British Steel Industry, p. 244Google Scholar.

⁵⁷ Temin, Peter, Iron and Steel in Nineteenth-Century America: An Economic Inquiry (Cambridge, Mass., 1964), pp. 69–93, passimGoogle Scholar; Hawke, Gary R., “The United States Tariff and Industrial Protection in the Late Nineteenth Century,” Economic History Review, 2nd ser., 27 (1975), 84–99Google Scholar.

⁵⁸ See, for instance, , Landes, The Unbound Prometheus, pp. 268–69Google Scholar.

⁵⁹ , Burnham and , Hoskins, Iron and Steel, pp. 272–76Google Scholar; , Hoffmann, Das Wachstum, pp. 454–55Google Scholar.

⁶⁰ Statistisches Jahrbuch, 7 (1886), 36 ffGoogle Scholar; Statistik des Deutschen Reichs, Vierteljahresheft (1914), pt. I, pp. 358–65Google Scholar; Kuczynski, Jürgen, Die Ceschichte der Lage der Arbeiter unter dem Kapitalismus, 38 vols. (Berlin, 1962–1970), Darstellung der Lage der Arbeiter in Deutschland von 1871 bis 1900, vol. 3, pp. 425 -26; Darstellung der Lage der Arbeiter in Deutschland von 1900 bis 1917/18, vol. 4, p. 427. Increases in labor productivity, due to improvements in technology or in the quality of labor, which surely occurred, would bias the elasticity estimate downward. This probably outweighed any upward bias due to downward shifts in the supply curve of labor because of better transportation and communicationsGoogle Scholar.

⁶¹ Formally, IG = [(C_g + C_row)/Q_g]≠ - [Q_row/Q_g]E where IG is the elasticity of demand facing the German industry, C_g and C_row are the quantities consumed by Germany and the rest of the world (excluding the U.S.A.), Q_g and Q_row are the quantities produced by Germany and the rest of the world, and ≠ and η are the underlying elasticities of demand and supply.

⁶² Stern, Robert M., Francis, Jonathan, and Schumacher, Bruce, eds., Price Elasticities in International Trade: An Annotated Bibliography (Toronto, 1976), p. 356Google Scholar.

⁶³ In 1910–14 C_g equalled 9.94 million tons; C_row equalled 30.30 million tons; Q_g equalled 14.79 million tons; and Q_row equalled 25.15 million tons. , Burnham and , Hoskins, Iron and Steel, pp. 275–79Google Scholar.

⁶⁴ , Carr and , Taplin, British Steel Industry, pp. 164, 230Google Scholar.