2. The Second Industrial Revolution and the State
a) A dramatic shift
b) Protection and cartels
c) How Tariffs and Cartels enhance Growth
d) Trade Policy
e) Cartels and Anti-Trust Policy
3. Background on Industrial Chemistry
a) Inorganic chemistry
b) Organic chemistry
4. Exceptional Growth
a) Size of the Industry
b) Inorganic Chemistry
c) Organic Chemistry
5. Market Structure
a) Firm Size and Business History
c) United Alkali
6. How the State promoted cartels
a) Legal Recognition
c) University Research
d) Patent Law
7. How Cartels produced growth
a) Economies of Scale and Scope
Within half a century before World War I, the economic global landscape changed dramatically. Britain remained the world’s largest exporter and dominant in many industries and London was still the centre of global financial markets and the organizer of world trade. But Germany and the U.S. were catching up dramatically. Not only both countries were rapidly gaining shares in world manufacturing and world trade and catching up in income p.c., they even took technological leadership in a number of key industries like steel production, electrical engineering and chemicals.
This paper provides a detailed analysis of one of these new industries, the chemical industry, in one of the follower countries, Germany. This may serve as a fruitful case study to understand the broader pattern, since here the success of German industry is especially marked. In the established inorganic (soda producing) chemical industry, Germany caught up to Britain that had dominated world markets for decades. Even more significantly, in the new organic (dyestuff producing) chemical industry, English firms were leading until the 1860s, but by 1913 Germany produced about 85% of the global dyestuff output. Further, the chemical industry highlights a feature of the German industrial structure in a drastic way that may have been decisive for their success: There was heavy market concentration in the form of oligopolies and cartels. While other countries reacted with anti-trust legislation, the German state promoted this development by legalizing cartels, imposing tariffs and – in the case of dye industry – the close cooperation of business and state universities in research. I will argue that this oligopolistic structure was a key factor for the development of the German chemical industry. In short, the argument goes as follows:
state action à market concentration à industrial growth
This paper tries to answer the four key question of this framework. Was there exceptional growth in the industry? Was there market concentration? Did public policy promote market concentration? Did market concentration promote industrial growth?
The argument is modeled after a paper by Steven Webb about the German steel industry. He argues that the state promoted cartelization of metallurgy through legal support and the tariff hike in 1880 that lowered foreign competition and made dumping possible. This was more successful in the heavy branch of the industry, where larger capital requirements due to economies of scale constituted a natural barrier for entrants. He argues that cartelization fostered efficiency through vertical integration, incentives to invest, and reduced uncertainty. Cartel contracts normally limited sales, not produced quantities. As a consequence, firms integrated forward into less cartelized industries to produce more without violating the cartel contract. They also integrated backward into cartelized industries to avoid paying the cartel mark-up. Vertical integration in steel industry was essential for an increase in productivity for two reasons. First, the iron didn’t have to be reheated when it was used to make steel. Second, several new energy-saving technologies were only possible in vertical integrated plants. Since the quotas were fixed based on capacities and renegotiated periodically, cartels provided an additional incentive for firms to invest. In addition, they stabilized prices and raised them moderately, reduced uncertainty and made large long-term investment less risky. The overall success of Germany iron and steel industry seems to support Webb’s argument. From 1879-1913 German steel production increased 10.4% p.a., 4.4% more than British production and more than three times as fast as the German economy. Her share in global steel market (excluding the US) rose from 22% to 39%. British steel production decreased 1893-1913 from 98% of the German production to 44%. While in the 1870s, Britain produced four times as much iron and twice as much steel as Germany, in 1912 Germany made twice as much iron and steel as Britain. In 1913 Germany was producing 16 million tons of steel, at least ⅔ of the production of entire Europe.
This paper confirms that Webb’s arguments are broadly applicable for the chemical industry as well. The industry was highly successful, was heavily concentrated, and the state supported concentration. Nevertheless, the details of the picture have to be changed. Tariffs played an important role for cartelization only in the inorganic branch of the chemical industry. On the other hand, the early establishment of state research facilities and their close cooperation with dye businesses laid the foundation for cartelization of the organic chemical industries. Vertical integration reduced transportation costs in dye production as sharply as in steel production. Stable and high prices were at least as decisive for chemical firms as for steel producers, since investments were huge, specific, and repaid only in the very long run. In contrast to steel cartels, chemical cartels did not induce higher investment in production facilities, since quotas were not renegotiated regularly based on production capacities. However, they induced the firms to diversify into new industries like pharmaceuticals, where large economies of scale and scope could be realized by using the same equipment and inputs as in dye production.
2. The Second Industrial Revolution and the State
a) A dramatic shift
The time of the so-called Second Industrial revolution, roughly from 1860 to 1913, witnessed the rise of Germany to a position of economic dominance in Europe. “To contemporaries this seemed the most outstanding fact of the period.” At the eve of World War I, Germany produced ⅔ of all steel in Europe, 20% more electrical energy than Britain, France, and Italy combined, more coal than the rest of the continental countries together, and consumed twice as much raw cotton as France. In 1890, German exports were 10% higher than in France, in 1911, they were 60% higher. A more systematic analysis confirms this somewhat narrative evidence. Measured in income p.c., most European countries lost ground relative to Britain. In contrast, American and German incomes rose ¼ to ⅓ compared to Britain (table 1). 1870-1913, real GDP p.c. grew in Germany faster than in any other European country. The picture is even more dramatic when looking at global manufacturing output. While Britain lost market shares, and most other countries rather stagnated or fell back even further, Germany and the U.S. gained rapidly and produced as much as ½ of the world’s manufactured goods in 1913 (table 2 and 3). A similar pattern can be found in levels of industrialization. While Britain’s index value rose 80% over the period and most countries’ industrialization grew about two-fold, German and American industrialization grew five-fold (table 4). While Britain’s GPD p.c. growth slowed down markedly after the mid-19th century, Germany’s and U.S. growth increased significantly (table 5, 6, and 7). This fast growth of income per capita took place despite fast population growth in Germany and America (table 8).
b) Protection and cartels
Development theories based on protection depend on the assumption that tariffs protect a certain industry. That means the tariffs increase the prices of the protected goods relative to other goods. This is all but trivial. First, besides tariffs, there are non-trade barriers like currency devaluations, quotas, voluntary export restrictions, and anti-dumping duties. Luckily, they were virtually unknown in the 19th century. Second, it is not easy to calculate the degree – or even the direction – of protection. Tariff laws in our period often specified specific tariffs. To calculate ad valorem tariffs, one needs information about the market prices of the goods. To calculate effective rates of protection information about imported inputs and their tariffs is needed. Finally, to calculate real or net protection, one needs to model the price changes of domestic inputs due to the tariff structure. While for the U.S. there is some data on effective rates of protection and net protection available, to my knowledge there has been no systematic attempt to calculate either one for Germany. Many authors use simple ad valorem or even specific tariffs to evaluate protection.
Webb argues that there is a strong causal relationship between tariffs and cartels. If an industry is not competitive on world markets, tariffs are necessary for cartels – without them there is always an outsider (the foreign competitor). If the industry is competitive on world markets, tariffs are necessary for dumping, by making re-imports more expensive. Dumping is an easy way to keep the domestic price stable (and high) while always producing at full capacity. So Webb claims that “the tariffs and cartels were the main pillars of a structure that restrained competition from both foreign and domestic sources.” Contemporaries both in Germany and the U.S. were very aware of the interdependence, as reflected in the famous expression that “the tariff is the mother of trusts”. Indeed, the public debate about the 1879 tariff hike in Germany was centered on cartels and dumping.
c) How Tariffs and Cartels enhance Growth
Economic theory provides six theories how protectionism might support growth. First, if tariffs create significant government income, interest rates fall and private investment is crowded in. Second, tariffs may induce a sectoral shift of resources from agriculture to manufacturing, where productivity growth is higher. Third, protectionism may induce tariff jumping, that is the import of production factors. Four, if tariffs are biased towards consumption goods, they make capital goods relatively cheaper and increase investment. Five, if there are static or dynamic scale economies or positive externalities of human capital accumulation or agglomeration effects, the infant industry argument applies. Six, if foreign markets are oligopolistic, tariffs may capture some of the rents and increase domestic welfare. An assessment of these direct effects is beyond the scope of this paper. Here tariffs are only interesting in their role as promoters of cartelization.
At least since Josef Schumpeter, economists have argued that large corporations, oligopolies, cartels or other forms of market concentration have a positive effect on technological progress and economic growth, in short, that “perfect competition is not only impossible but inferior.“ Even in a static model some market concentration is efficient, if there are large scale economies and a limited market size. But Schumpeter goes further and argues that only in the absence of perfect competition are firms able to do large-scale investments and conduct research and development. Three reasons are important. First, if there are credit constraints, monopoly rents are necessary to finance future investments ex ante. Second, if firms are risk averse, stabilizing the price increases investment. This observation is even more important in a world full of “Keynesian” uncertainty. Third, and most fundamentally, monopoly rents are the only incentive for innovation in market economies. Even without uncertainty and credit-constraint a risk-neutral firm will not invest in innovation if there are not some barriers to entry that protect their profits. Any innovation has to be financed ex post by some form of monopoly rent. Webb adds a fourth argument to the list: Cartels promoted vertical integration and vertical integration increases productivity.
In addition, Schumpeter argues that neoclassical theory not only neglects the gains of market concentration, but overstates its social costs dramatically. This is because the main the form of competition is based on technological change (Creative Destruction) – and not price competition. But as many later researchers have pointed out, the relationship between market power and growth is not that linear and simple. If market concentration is good or bad for innovation and growth depends on additional factors like industry characteristics and economical and political institutions. Ceteris paribus, market concentration will be better for growth if there are economies of scale and firms have to invest large sums in capital or R&D. I will show that this was the case in Germany’s chemical industry and indeed market concentration here promoted growth.
d) Trade Policy
This section provides some historical background on the development of trade policy and policies on monopolies and cartels. Between the Napoleonic Wars and World War I, Europe experienced the fastest growth of trade in goods in her history. The development of protectionism can be imagined graphically as a U-shaped curve with high degrees of protectionism at the beginning and the end of the period and most liberal policies around 1870. During a first phase (1815-46), Britain gradually opened up and abolition of the Corn Law in 1846 is commonly seen as the beginning of the free trade era of the U.K. A second phase (1846-60) witnessed the spread of free trade on the Continent, most notably the internal integration of the German market. In 1860, the Cobden Treaty between Britain and France opened the third phase (1860-79), when tariffs between European states were reduced significantly. This was done by a number of temporarily limited trade agreements with most favored nation clauses. During the fourth phase (1879-92), the Continent returned to protectionism. Beginning with Germany in 1880, all major countries imposed higher tariffs during the following 13 years. Only Britain and Denmark kept their markets open. The historical trade literature sees pressure from farmers as the main reason for the tariff hike. Both the dramatic reduction of transportation costs due to railroadization and increases of agricultural output in land abundant countries due to immigration and mechanization reduced world market prices (table 12). While imports to Europe from North America, Ukraine, and India increased dramatically, the productivity of European agriculture slowed down. Note that it is not clear a priori if the slow down was a cause or a consequence of the imports. Anyways, agricultural producers lobbied successfully for protection. During a fifth phase (1892-1913), protectionism was strengthened on the continent while Britain remained a free trader.
Germany overtook Britain in terms of exports to European countries in the 1880s and by 1913 she had caught up with Britain in terms of overall exports. Germany’s exports grew rapidly both in periods of liberalization, free trade, and increasing protectionism. As late as 1818, internal tariffs within Prussia were abolished and during the 1820 several other states have founded regional free trade areas. The Zollverein, a customs union of almost all German states, was founded in 1834. Since Prussia, the economically and politically dominant force in the Zollverein, was a traditional grain exporter, and the Junkers were a highly influential class in Prussia, the general tariff policy of the Zollverein was rather liberal. During the 1860s, Germany became a net importer of agricultural goods. Grain imports doubled during the 1860s and doubled again during the 1870s, in total from 0.7 million tons to 3 million tons annually. At the same time, growth of agricultural output slowed down from 2.5% p.a. to 0.4% p.a. In addition, industry suffered when the Gründerboom after the Franco-Prussian war was followed by the Gründerkrise, the strongest downturn in 50 years. This sharp crisis was the beginning of a 20 year period of sluggish economic growth and falling prices, often called the “First Great Depression.” Pressure on prices of both agricultural goods and industrial goods during Grain Invasion and Great Depression set the stage for the “alliance of rye and iron” and the wave of protectionism that arrived a few years later.
The tariff law of 1879 (in effect in 1880) increased import duties significantly for industrial goods (often to 15%-25% in ad valorem terms) and only moderately for grains (about 5%). In 1885, 1887, 1902 and 1906 tariffs on grains were raised to more than 34% while other tariffs remained fairly stable. But the tariffs were still lower than before 1860 and lower than in France, Russia, or the U.S. The tariffs were imposed for three reasons. First, both the rural Junker class and capitalists lobbied to receive protection from falling world market prices. Second, Erziehungszölle should promote infant industries. Third, since the Reich relied heavily on direct transfers from the states (Matrikularbeiträge) and could not impose direct taxes, the tariff hike was supposed to increase the fiscal autonomy of Chancellor Bismarck. As mentioned above, a rigorous account of the effects of the tariff on relative prices is still lacking, so we cannot evaluate if the tariffs had the intended effects on individual sectors and industries. But industry studies have shown that for example the iron and steel enjoyed positive effective rates of protection. However, since after the 1890s the German steel industry was highly competitive on world markets, the persistence of the tariff makes only sense when one takes into account its effect on cartels.
For our period there is a strong and robust statistical correlation between tariffs and growth on a global level. This was argued by historians for a long time, but has been recently confirmed by more rigorous statistical analysis that controls for other variables. This pattern seems to hold for both cross-country and time series data.
e) Cartels and Anti-Trust Policy
The different German business associations like “interest groups,“ “pools,“ “rings,“ “trusts,” “syndicates,” “cartels,“ “unions of employers” can be understood as three different types of associations: organizations for collective bargaining with trade unions (Arbeitgeberverbände); lobby groups (Interessenverbände), and cartels. It is estimated that in 1911 there were 600 cartels in Germany and in 1926 about 3000. According to a standard economic definition, cartels are an “associations of legally independent firms with the intent or effect of influencing the market by means of regulation of price or quantity produced.” This was done by i) allocating regions or individual customers between producers, ii) fixing prices, iii) fixing volumes of sale, or iv) sharing the profits according to fixed quota. But German cartels in our period went further and often can best understood as “partial integration,“ an intermediate type of organization between a textbook cartel, a joint venture and a merger. For example, cartels set up single sales agencies for their members, coordinated research and development, bought or set up common subsidiaries to produce inputs, or issued bonds on the London capital market when its members where to small individually. It is important to keep in mind that in the context of this paper the term “cartels” is used in the same way, that is to describe all these forms of non-market coordination between competitors.
A handful of early cartels was established in German states after the 1820s. While the number of lobby groups increased already in the 1850s, cartels were virtually unknown in the 1870s, increased in number after 1880 and developed most rapidly from the 1890s onwards. A common argument in the literature is that the crisis of the 1870s led to cartelization, but most cartels were founded after the serve recession of the Gründerkrise (1873-4). Webb argues that cartels developed successfully only after the new tariff made re-imports impossible and lowered competition from foreign firms. But he lacks a convincing explanation why it took often another ten to twenty years until major or even national cartels were established.
The three classical examples of German cartels are in the steel, coal, and potash industries. A successful rail cartel was established in 1859, but collapsed during the Gründerboom. During the 1880s several smaller cartels were established and in 1899 an umbrella cartel of pig iron producers was founded that covered whole West Germany. But only in 1904 a cartel of all major producers of heavy steel products were founded (Stahlwerksvereinigung). While Webb claims the new tariff was the reason for the successful cartelization, others have argued that the steel producers reacted to the monopsony that was evident on the rail market after the nationalization of the railroads in the 1870s. Although plans for an outright merger into a single trust had been made since 1900, it took until after the War to found the Vereinigte Stahlwerke. In coal mining the Rheinisch-Westfälisches Kohlesyndikat was set up in 1893. It controlled 84% of the Ruhr coal production and in 1904 a common unit for marketing and sale was established. Five years later the state-owned Prussian mines became cartel members. During WWI the state made membership de facto compulsory. The potash mining industry in Strassfurt had a world monopoly. In 1879 a cartel was set up, both by private and state-owned mines, the latter holding about 8% of the sales quota. In 1910 the cartel broke down and the state set up a compulsory cartel by law (Reichskaligestz). Since Germany had a global monopoly export prices were usually held higher than domestic prices – contrary to the normal cartel behavior of dumping.
Four mechanisms were important to stabilize the cartels. First, “natural” barriers in the form of large capital requirements made the entrance of outsiders in heavy industries more expensive. Second, common sales organizations made monitoring easier. Third, the system of German universal banking helped. Banks who held shares and sent members to the firm’s supervisory boards could pressure on firms to join cartels, as famously happened in the case of the Phoenix steel producers. They further helped monitoring since they often had intimate knowledge about sales and economic transactions, e.g. when railroad companies financed their purchases of rails with bank credits. Notably, a number of important cartel contracts were negotiated under mediation of bankers and signed in banks. Finally, the state supported cartelization in several ways, as discussed in section 6.
3. Background on Industrial Chemistry
In our period, the chemical industry consisted of two branches that varied significantly in their maturity, products, technical processes, and industrial structure.
a) Inorganic chemistry
Inorganic chemistry, or “heavy” or “mineral” chemistry, was a quite old and well established industry. The main product of the branch was soda (alkali; sodium carbonate). In addition, sulphuric acid, hydrochloride acid and bleaching powder were produced. Soda was mainly used for cleaning, dyeing and bleaching of textile, soap, glass, and iron. Later it was also used as an intermediate product for organic chemicals. In 1791 Nicholas Leblanc of France gave birth to the chemical industry by developing a process of soda synthesizing. The Leblanc soda process converts chalk (calcium carbonate), limestone (calcium carbonate), salt (sodium chloride), and coal to soda and hydrochloride gas. For 80 years Leblanc’s technology dominated the rapidly increasing soda production. Soda was produced in England and to a lesser degree in France, since here the Industrial Revolution had created a large demand that could not be satisfied by natural substitutes like potash. Soda was produced in the form of soda ash, soda crystals, and caustic soda. In the 1860s, the Solvay family from Belgium built the first soda plant using the ammonia-soda process, and within 30 years this technology had replaced the Leblanc process. The Solvay process converts salt with ammonia and carbon dioxide to ammonium chloride and soda. The new process was cheaper and produced soda of higher purity. The world soda production increased from 0,15 million tons in 1863, all produced using the Leblanc process, to 1,8 million tons in 1902, 82% produced by ammonia-soda process and soda prices had fallen by two thirds. Leblanc producers could survive quite a long time since they produced hydrochloride acid as a by-product. In the form of chlorine it was used after the 1770s for bleaching of textiles. As soon as hydrochloride acid could be produced by electrolysis, Leblanc soda manufacturing lost its last source of competitiveness.
 Webb (1980): Tariffs, Cartels, Technology.
 Today “dumping” refers often to firms that sell under costs to drive out competitors. Originally (and I use it that way) dumping meant selling abroad for a lower price than at home. Without tariffs, re-imports make dumping impossible.
 Very valuable background information about the process of iron- and steelmaking, technical improvements and their history, including pictures, graphics, and references, can be found in both German and English language on Wikipedia.
 For a discussion of the case of potash and coal mining see Kleeberg (2007): German Cartels, p. 15, 27. For a somewhat supporting view see Bloch (1932): On German Cartels. He argues in the context of the Great Depression of the 1930s that the strong incentives for building-up capacity caused tremendous excess capacity in cartelized industries. Note that other authors have been very critically regarding the dynamic efficiencies of concentrated markets in the steel industry and the ability of large firms to innovate. See, for example, Adams & Dirlam (1966): Big Steel, Invention, and Innovation.
 Typically they could raise prices in the range of about 10%. Webb (1980): Tariffs, Cartels, Technology, p. 311.
 A similar argument was made for the coal industry by Henderson (1975): German Industrial Power, p. 183. See also Bloch (1932): On German Cartels.
 Neither for the iron and steel nor for the chemical industry at the turn of the 20th century there exist a careful micro analysis of the reasons of cross country variation in productivity. The only industry study I am aware of is Gregory Clarks (1987): Why isn’t the Whole World Developed?; a study of the textile industry. He finds that “local culture” explains the largest part of variation in productivity between countries. Both Webb and this text have to rely more on indirect evidence.
 Webb (1980): Tariffs, Cartels, Technology, p. 327.
 Landes (1972): Unbound Prometheus, p. 269.
 Milward & Saul (1973): Development of Continental Europe, p. 19.
 Milward & Saul (1973): Development of Continental Europe, p.17.
 Milward & Saul (1973): Development of Continental Europe, p.19-20, 60.
 Williamson (1998): Globalization and Policy Backslash.
 This is not true for Scandinavian countries, which experienced fast growth over the period, as well as some recently settled countries like Argentina and Canada. Russia also grew pretty fast throughout the period, but probably much slower than Germany.
 For Europe see Bairoch (1989): European Trade Policy, p. 72; for the U.S. see Irwin (2006): Tariff in Americas Gilded Age, p. 1-4.
 For an exhaustive overview see Corden (1971): Theory of Protection; for a short introduction see Irwin (2006): Tariff in Americas Gilded Age, p. 3-12.
 For effective rates of protection see Hawke (1975): U.S. Tariff and Protection; for relative prices see Irwin (2006): Tariff in Americas Gilded Age. Webb (1980): Tariffs, Cartels, Technology gives ERP for different branches of metallurgy.
 Webb (1980): Tariffs, Cartels, Technology, p. 312.
 For contemporary and historical discussions see McVey (1899): Trusts and the Tariff; Clarke (1907): Relation of Tariff to the Trust; and Kronstein (1942): Cartels and Patents. The statement is by Henry Havemayer.
 Kleeberg (2007): German Cartels, p. 12.
 This argument was – and still is – the argument most often used by debates by economists as well as non-economists. Friedrich List argued famously in the 1840s – and his arguments were broadly discussed in the late 19th century – that less developed countries needed “educational tariffs” (Erziehungszölle) to compete against Britain until scale economies and externalities make the domestic industry competitive.
 For a critical assessment of these arguments in context of late 19th century industrial countries see Broadberry (1998): How did the U.S. and Germany overtake Britain?; Irwin (2000): Tariffs and Growth in Late 19th Century America, p. 17; Krause & Puffert (2000): Chemicals, strategy, and tariffs.
 Schumpeter (1942): Capitalism, Socialism, and Democracy, p. 106.
 Keynes (1937): The General Theory of Employment; Crotty (1996): Theory of Investment.
 This is Schumpeter’s main argument.
 For an excellent survey of the theoretical and empirical literature, see Scherer (1992): Schumpeter and Plausible Capitalism.
 During the 18th century, trade in goods within Europe increased no more than 2-3 fold, from 1815-1914 40-fold and from 1914-84 12-fold. The export share of GDP in Europe increased from 3% to 12%. Bairoch (1989): European Trade Policy, p. 1.
 The following periodization is from Bairoch (1989): European Trade Policy, the standard text about the topic.
 In 1846 tariffs on almost all goods, not only agricultural goods were reduced. But this didn’t mean free trade: Tariffs on wheat, for example, stayed at 20% until 1869. Similar “free trade” in Germany and “free trade agreements” like the Cobden treaties did not mean zero tariffs, but a significant reduction.
 Even after the major war ot the period in 1871, France and Prussia signed immediately a even more far-reaching free trade agreement.
 U.S. export of wheat increased from the 1840s to the 1870s from 50.000 tons to 2.1 million tons p.a. (40 fold or 13% p.a.) while wheat imports rose to European countries rose from 0.3%-6% of domestic consumption in the 1850s to 19%-100% in the 1890s. Bairoch (1989): European Trade Policy, p. 47.
 On the Continent, growth of agricultural production declined from 0.4-0.5% p.a. p.c. to -0.2%. Bairoch (1989): European Trade Policy, p. 46.
 High imports may reduced domestic prices and reduced incentives so that productivity stagnated. On the other hand, stagnating productivity could have caused excess demand for foodstuffs that were satisfied by imports. Paul Bairoch is very clear here (without proving his argument): The slowdown in productivity growth is “almost completely explained by the influx of oversees grain, itself the result of the drop in transportation cost, and of the total abolition of tariff protection.”
 Williamson (1998): Globalization and Policy Backslash, p. 66; O’Rourke (2000): Tariffs and Growth, p. 458, Henning (1973): Industrialisierung in Deutschland, p. 227.
 Bairoch (1989): European Trade Policy, p. 72-3.
 Bairoch (1989): European Trade Policy, p. 15-7.
 Bairoch (1989): European Trade Policy, p. 58, 61.
 For an impression see figure 60 (prices of investment goods) in Henning (1973): Industrialisierung in Deutschland, p. 274. Building of new railroads increased fourfold between 1868 and the peak of the crisis (from 670 km to 2400 km) and then slumped to 1000 km in 1878. See Henning (1973): Industrialisierung in Deutschland, p. 205.
 For example, pig iron prices fell from almost 150 marks per ton to about 50 marks in the mid 1880s and recovered afterwards. Henning (1973): Industrialisierung in Deutschland, p. 211; See the overview of data in Tilly (1991), table 9.2, and p. 180. For example, the peak-to-peak average growth rate of the business cycle of 1874-84 was one half to one third of all other business cycles from 1857 to WWI. For an interpretation as the “Great Deflation” see Henderson (1975): German Industrial Power, p. 176.
 Henning (1973): Industrialisierung in Deutschland, p. 203-36.
 Baldwin & Martin (1999): Two Waves of Globalization, p. 26.
 In 1890 about ⅔ of all central government income came from tariffs. The situation was very similar in the US. Irwin (2002): Interpreting the Tariff-Growth Correlation, p. 15, figure 3.
 Webb (1980): Tariffs, Cartels, Technology.
 O’Rourke (2000): Tariffs and Growth; Irwin (2001): Tariffs and Growth; Irwin (2002): Interpreting the Tariff-Growth Correlation; Irwin (2002): Did Import Substitution promote Growth? Irwin’s articles are largely redundant.
 Bairoch (1989): European Trade Policy.
 Kleeberg (2007): German Cartels, p. 2.
 For a general discussion of the character of cartels until the Great Depression see Bloch (1932): On German Cartels and Kronstein (1942): Cartels and Patents. Kronsteins gives a similar definition on p. 644, footnote 1.
 These are the “four types of cartels” in Henderson (1975): German Industrial Power, p. 179-80.
 The only cartel that borrowed in its own name was the potash cartel. Kleeberg (2007): German Cartels, p. 17.
 Kleeberg (2007): German Cartels, p. 10.
 1870 there have been six cartels; during the Gründerkrise a couple cartels emerged, but of the 275 cartels in operation in 1900, nearly 200 had been founded between 1879 and 1890. See Henderson (1975): German Industrial Power, p. 179.
 For example see Kleeberg (2007): German Cartels, p. 19.
 For a detailed discussion of cartels in the steel industry, see Webb (1980): Tariffs, Cartels, Technology.
 Kleeberg (2007): German Cartels, p. 12-3. For an account of the German railroad history see Henning (1973): Industrialisierung in Deutschland, p. 238-243.
 For a detailed discussion of cartels in coal mining, see Henderson (1975): German Industrial Power, p. 182-3; and Kleeberg (2007): German Cartels, p. 17-31.
 This is an example of a broader pattern: During the WWI and the interwar period, the state used threats of compulsory cartels more and more to enforce private cartelization (to terms of the state). See Kleeberg (2007): German Cartels, p. 31.
 For detailed discussions of the compulsory potash cartel see Haber (1971): 1900-1930, p. 118-120; and Kleeberg (2007): German Cartels, p. 14.-17.
 Gerschenkron (1962): Backwardness, p. 15; Edwards & Sheilagh (1996): Universal Banks; Kleeberg (2007): German Cartels, p. 40-8.
 There is only an indirect connection between sodium carbonate and carbonized soft drinks, although both are referred at as “soda”. Sodium carbonate can be used to produce carbon dioxide, which is dissolved in water to produce carbonized water. Baking soda, in contrast, is sodium bicarbonate and thus chemically closely related to sodium carbonate.
 Teltschick (1992): Deutsche Großchemie, p. 16-7.
 For a detailed analysis see Haber (1969): Nineteenth Century, p. 100-2.
 Aftalion (2001): Chemical Industry, p. 59.
 Up to that date, textiles were treated with soda, dipped into buttermilk and then exposed to sunlight. This was repeated many times. Using chlorine reduced the time for bleaching from several weeks to few hours and also reduced land consumption enormously.