Jump to ContentJump to Main Navigation
The Rise of “The Rest”Challenges to the West From Late-Industrializing Economies$

Alice Amsden

Print publication date: 2001

Print ISBN-13: 9780195139693

Published to Oxford Scholarship Online: November 2003

DOI: 10.1093/0195139690.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2019. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.  Subscriber: null; date: 17 September 2019

Tribulations of Technology Transfer

Tribulations of Technology Transfer

Chapter:
(p.51) 3 Tribulations of Technology Transfer
Source:
The Rise of “The Rest”
Author(s):

Alice H. Amsden (Contributor Webpage)

Publisher:
Oxford University Press
DOI:10.1093/0195139690.003.0003

Abstract and Keywords

In theory, technology transfer should enable a backward country to achieve world productivity norms, but in practice, because technology is ‘tacit’, and never completely codifiable, the best technology transfer rarely achieves productivity parity between buyer and seller. The more tacit a technology is, the more difficult it is to transfer, and the more monopolistic the power of the seller and the lower the skills and organizational capabilities of the buyer, the worse the transfer. A shyness of foreign investors left the successful late industrializing countries (the rest) with a serious skill deficit that grew over time relative to that of the North Atlantic and Japan. The tacitness problem arose early because of the sectoral composition of the manufacturing output of the rest; whatever the source of manufacturing experience, all countries tended to share the same sequential industry mix based on natural resources, and because the specific properties of a natural resource vary by location, a successful technology transfer requires substantial investments in local learning and adaptation. Japan set a benchmark for learners that began with technology transfer; it started to industrialize rapidly only in the 1890s, at about the same time as China, and slightly after Brazil, India, and Mexico, but, given its engineering capabilities and basic knowledge, its absorption of foreign know‐how was more proactive, systematic, and thoroughgoing than that of other countries, and is analysed below in the case of the silk industry.

Keywords:   Japan, investment, late industrialization, newly industrialized countries, North Atlantic, silk industry, skill deficit, tacitness, technology transfer

In theory, technology transfer should enable a backward country to achieve world productivity norms. In practice, because technology is ‘tacit’ and never completely codifiable, the best technology transfer rarely achieves productivity parity between buyer and seller.1 The more tacit a technology is, the more difficult it is to transfer. Given any level of tacitness, the more monopolistic the power of the seller and the lower the skills and organizational capabilities of the buyer, the worse the transfer. Transfer may be most perfect through the medium of direct foreign investment, but such investment does not necessarily arrive when it is needed.2 If it arrives too early, it may “crowd out” national firms (see chapter 8).

A shyness of foreign investors left “the rest” with a serious skill deficit that grew over time relative to that of the North Atlantic and Japan. The tacitness problem arose early because of the sectoral composition of “the rest's” manufacturing output. Whatever the source of manufacturing experience, all countries tended to share the same sequential industry mix. By the 1930s or much earlier, food processing (including tobacco and beverages) dominated at around 30–40 percent of total manufacturing output (especially the refining of sugar, brewing of beer, and milling of flour). Next in importance were textiles and apparel (cotton and silk).3 Then came cement, paper, matches, and, after the turn of the century, steel. These are all natural resource‐based industries.4 Because the specific properties of a natural resource vary by location, a successful technology transfer requires substantial investments in local learning and adaptation.

Japan set a benchmark for learners that began with technology transfer. Japan started to industrialize rapidly only in the 1890s, about the same time as China and slightly after Brazil, India, and Mexico.5 Japan's first Western style raw silk factory using imported equipment was founded by the Maebashi (p.52) local government as late as 1870 (the first modern Chinese silk filature, founded by a silk merchant, made its appearance in Shanghai in 1881 [Eng 1984]). Japan's pioneering firm in cotton textiles, the Osaka Cotton Spinning Company, was established with the assistance of foreign engineers in May 1882 (Miyamoto 1988) (the first modern cotton spinning mill appeared in China in 1890). Given Japan's engineering capabilities and basic knowledge, its absorption of foreign know‐how was more proactive, systematic, and thoroughgoing, as analyzed below in the case of the silk industry.

Technology Transfer

A technology transfer was always a necessary condition for late industrialization but almost never a sufficient one. Transfers were especially problematic before World War II, when transportation and communication were relatively poor and “the rest” was in an early phase of industrial transformation. Therefore, the process of technology transfer was probably less satisfactory in the period 1850–1950 than in the next fifty years.

Before the 1910s, foreign “firms” were less likely to establish manufacturing operations in “the rest” then were foreign “individuals.” Not a single Lancashire textile‐based company before 1910 operated as a multinational enterprise in China, India, Turkey, Mexico, or Brazil.6 The performance of an individual, moreover, was not necessarily comparable to that of a firm: “British individuals abroad are not equivalent to British managers, who can draw on the home company's on‐going experience and remain part of a business organization with knowledge of all facets of the business operations. Likewise, British trading companies that provided managerial supervision (management contracts) are not identical to a managerial organization that grows from an operating parent enterprise's experience in textile manufacturing” (Wilkins 1987, p. 121). The same weaknesses were true of the French, Spanish, and American individuals who, as noted in the last chapter, dominated Mexico's modern textile industry beginning in the 1830s. Italian immigrants often became entrepreneurs in Argentina, but a subsidiary of an Italian multinational, Pirelli, was first established in Argentina only in 1917 (Ines Barbero 1990). This was the same year that Ford Motor Company began assembling cars in Argentina (Diaz Alejandro 1970). Because individuals were not part of ongoing business organizations, their know‐how was not necessarily upto‐date. British engineers in India's cotton mills, for example, often followed technological developments in Lancashire with a considerable time lag (Kiyokawa 1983). Not infrequently, foreign know‐how bordered on the sham. In the 1880s, several Anglo‐Brazilian sugar factories were promoted by railroad contractors and were universally a failure: “Contemporary opinion was unanimous in regarding the direction of these companies as deplorable,” although (p.53) the financial success of some Brazilian sugar factories suggested that it was possible to run them successfully (Graham 1968, p. 153).7 Even the big trading companies, which preceded the multinationals, were not especially effective in their technology transfer. In China, two British silk mills were brought into being by prominent British trading companies, Jardine, Matheson & Company and Kungping Company, but “owing to poor management these filatures all failed within a few years” (Lieu 1936, p. 34).

Sometimes the effectiveness of the foreign technicians was constrained by cultural and social disparities. In the case of the Ottoman Empire in the 1850s, “Christian Europeans simply were not the most effective role models and were unpersuasive as opinion leaders, even in those instances when they knew the language. Their advice often was ignored. In many cases, the hired technicians believed their job was to run the equipment and not necessarily to teach new skills. The enormous wage differentials between foreign and Ottoman workers that were typical contributed to poor relations between the two groups” (Quataert 1992, p. 32). In Brazil before the abolition of slavery in 1888, the St. John del Rey gold mining company “simply adapted to local conditions. The British management of the St. John, for example, desperately wanted to employ free Brazilian labor and could not. They turned reluctantly [but profitably] to slave labor,” which extended the life of slavery and delayed the rise of modern industry (Eakin 1989, 266–67).

When foreign firms finally superseded foreign individuals as technology providers in “the rest,” they were more likely to enter a foreign market to enjoy an ongoing process rather than to be a first‐mover and act as a catalyst for industrial expansion.8 “Direct British investment in Brazilian manufacturing grew with Brazil's over‐all industrial power”, it did not lead it. Thus, “the British were not the cause of Brazilian industrialization. [Indeed], the activities of some of them tended to hinder that process. But others worked shoulder‐to‐shoulder with Brazilians to bring it about” (Graham 1968, p. 142). In the case of Mexican railways, “local companies constructed a total of 226 kilometers of track before North American capital arrived to construct the country's two major arteries” (Coatsworth 1981, p. 38).9 Ultimately, American and European multinationals invested heavily in the manufacturing industries of Latin America, particularly in consumer goods, but when they did so in large numbers beginning in the 1910s or 1920s, many modern industries had already been founded (Phelps 1936). Most Latin American cigarette firms were established in the early years of the twentieth century and some in the 1890s. They grew rapidly in Argentina, Brazil, Chile, and Mexico. In these markets, the largest in the region, British‐American Tobacco gained a beachhead either just prior to or after the First World War frequently by acquiring a local firm (Shepherd 1989). The founders of Argentina's meat packing industry included one British firm as well as two native‐owned firms, which were taken over in 1907 by American packers (Crossley and Greenhill 1977). (p.54) The Corning Glassworks and the Pittsburgh Glass Company bought controlling interest in Argentina's financially strapped Cristalerías Rigolleau Company in 1942, thereby acquiring “an old and prestigious firm that already enjoyed a commanding position in its field and established connections with both suppliers and buyers” (Lewis 1990, p. 51).

In China, except for a couple of unsuccessful attempts, no textile mill owned by a Westerner was established until 1914 whereas modern Chinese mills began appearing in the 1890s (Chao 1975). Some Japanese investments in China's cotton mills involved big zaibatsu as opposed to individuals, but whoever the investor, it took over existing Chinese mills; the Chinese themselves were the trailblazers (Koh 1966). Foreign firms invested in Chinese industries other than textiles, but such firms initially tended to be very small, with no notable names of multinational manufacturers among them (Allen and Donnithorne 1954).10 Foreign individuals in India were responsible for starting the jute industry, a major nineteenth‐century exporter. The initiative for railroad construction also came from foreigners. But Indians took the lead in creating the cotton textile, power generation, shipping, construction, sugar, iron and steel, engineering, agricultural implements, and later chemical, automobile, and aircraft industries (Agarwala 1986). Initially, London would not allow India to develop its own steel industry for fear that it would displace British steel exports to India. When such exports were challenged by German steel, a domestic steel industry became acceptable. The British “must have thought that the abolition of the irksome prospecting laws would induce English entrepreneurs to set up steel plants in India. However, only one Englishman made a feeble attempt to enter the field,” and India's first steel mill was built by a prominent Indian entrepreneurial family, the Tatas (Tripathi and Mehta 1990, p. 61).

In Turkey, the “foreigners” who often established modern production facilities were really émigrés who had lived in the Ottoman Empire for generations. For example, the largest textile factory built in Izmir before 1912–13 was owned by a descendent of old French and English commercial families active in the Izmir region since the late eighteenth and early nineteenth centuries (Clark 1969). Ironically, foreign investment in Turkey became real only after native non‐Muslims were driven out of the country after World War I by nationalists who hoped to create a larger economic role for native Muslim capitalists. Instead, foreign investors filled the breech and eventually accounted for 63 percent of manufacturing output (Keyder 1994).

As in “the rest,” so too in Japan: foreign investors were not the first‐movers. As late as the period from 1896 to World War I, “when the Japanese had already demonstrated their general progressive drive and their specific industrial aptitudes, direct foreign investments in manufacturing began to appear” (Reubens 1955, p. 220).11

(p.55) In theory, foreign firms are desirable because they provide “spillovers” (discussed in chapter 9) and a positive role model: “One cannot go into the Chinese‐owned (textile) mills in China (circa 1930) without realizing the influence of the Japanese‐owned mills” (Moser 1930, p. 66). Nevertheless, foreign role models may crush domestic competition. In China's cigarette industry, British American Tobacco (BAT), a giant multinational, and Nanyang, a local firm, competed head‐on in the 1910s for China's growing market. Chien Chao‐nan, the owner of Nanyang, put a deposit on a warehouse in the foreign concession area of Shanghai to begin production (Nanyang had accumulated experience producing cigarettes in Hong Kong using Japanese technology). “The very next day a BAT comprador tried to buy the building,” which started a vehement argument that only ended when one of BAT's own compradors (a Cantonese like Chien) “defended Nanyang's position and urged BAT's management not to force Chien to surrender his rights to the building.” Chien installed 119 American cigarette‐making machines and later bought the site (Cochran 1980, p. 74).12 In another case in the 1890s, entrepreneurs who attempted to manufacture textiles in the Ottoman Empire for local consumption (in Egypt) were brought to bankruptcy by the pressure of lobbyists for Manchester textile interests. The British ambassador first attempted to block the mill's construction with administrative delays but then, to insure his own reappointment against threats from English textile manufacturers, acted more vigorously in getting the local government to impose high production taxes on the mill. Construction was halted (Clark 1969).

Using a foreign firm as a benchmark was especially important in industries whose technology was changing rapidly, such as textiles (in the late nineteenth century the spinning ring was displacing the mule and automatic looms were gaining ground). In these industries, an engineering orientation on the part of management was essential to keep abreast of technological change. Yet technological expertise was not necessarily a characteristic of foreign investors. Japan's first major steel works received technical assistance from Germany in 1897, but “the German engineers did not work as hard as the Yawata Works had expected. They lacked the basic knowledge and abilities to lead Japanese engineers and foremen.” This was in spite of the fact that the chief engineer, Mr. Toppe, earned a very high salary—twice as much as the Prime Minister of Japan! Yawata reached the conclusion that “the German engineers who came to the Far East (at the turn of the century) were hardly first rate” (Yonekura 1994, p. 43).13 In Mexico, financing of industry “fell to a relatively small clique of (European) merchant‐financiers who, because of their backgrounds in commerce and money‐lending, were more adept at rigging the market and manipulating government policy than at streamlining production methods or innovating new processes or techniques” (Haber 1989, p. 5). Foreign firms accounted for roughly 20 percent of output in India's textile industry, but they were hardly exemplary models. As (p.56) table 3.1 indicates, few directors either in European‐owned mills or Indian owned mills had a technical background; mercantile backgrounds were the norm in both cases.

Follow‐up investments by technology buyers were necessary to adopt and modify foreign imports and to absorb foreign know‐how. This was especially so in “the rest” because most of the industries being opened there before World War I were heavily raw material‐based, as noted earlier, and local investments were necessary to adapt foreign techniques to local processing specifications. British‐American Tobacco, for example, invested heavily in training Chinese farmers to grow American‐type tobacco (Cochran 1980). The scholar‐comprador Cheng Kuan‐ying organized, developed, and then managed the Shanghai Cotton Cloth Mill beginning in 1879. To determine whether foreign machinery could actually process China's raw cotton, Cheng contacted a fellow‐Cantonese in the United States and asked him to engage a technical expert to investigate the matter. The American, A. W. Danforth, came to Shanghai and expressed doubts about the suitability of the shorter fibers of Chinese cotton for machining. Cheng then sent Danforth back to the United States with some samples of raw Chinese cotton, and the resulting cloth was found to be equal in quality to American cloth (Hao 1970). (Production was delayed, however, because Cheng allegedly concentrated power in his own hands and “tended to treat the company's funds as his own” [Feuerwerker 1958, p. 212].) In the case of steel making, cost and quality are highly sensitive to the right choice and combination of raw materials, which cannot be determined simply by a mathematical formula. “They require patient and careful full‐scale experimentation, which means years of painstaking effort to determine the best combinations.”14 Thus, Japan's Yawata Steel Works was in big trouble initially because “the blast furnace designed by a German engineer was too large for the soft and low‐quality Japanese coke. Because of the high pressure and friction of the large furnace, the Japanese coke was crushed into powder and prevented air circulation and down flow of iron ore and flux” (Yonekura 1994, p. 44).15

Table 3.1. Background of Indian Factory Directors (Cotton Textiles), 1913 and 1925

1913

1925

Directors

Total

Technically Trained

Total

Technically Trained

European

30

4

24

2

Indian

132

8

151

9

Source: Adapted from Kiyokawa (1983).

(p.57) Just as teaching in technology transfers was far from ideal, learning was also imperfect due to insufficient local investments to absorb foreign knowhow. In 1890 as much as 60 percent of all technical personnel in the middle management of Bombay textile mills was European. As late as the 1920s, roughly one third of all such managers remained foreign (Kiyokawa 1983). Insufficient learning by Indians to dispense with the services of foreign advisers had apparently occurred. The Mexican textile industry may have started in the 1830s, but in the 1890s “foreign visitors commented that plants were managed by an Englishman with sound Lancashire experience or by men trained in the Manchester district of England. In 1896 a new plant in Torreon brought in forty skilled workers from France” (Keremitsis 1987, p. 197). One of Brazil's largest cotton mills, America Fabril, was started by two merchants and an industrialist in 1878. But as late as 1921, its managing director was a Yorkshireman and more than forty English foremen were engaged in various departments (Pearse 1929). A visitor to Sao Paulo in 1930 “found foreign technicians very numerous; one sees them everywhere” (Dean 1969, p. 177). The Shanghai Cotton Cloth Mill, as noted earlier, first used the American A. W. Danforth as a consultant. Later, after a fire, Danforth “was entrusted with the technical responsibility for erecting a plant, purchasing machinery, and organizing production for the new Hua‐sheng Mill” (Feuerwerker 1958, p. 221). Apparently, local managers in the first plant expansion had not acquired the capabilities (or trust) to undertake these tasks themselves in the second plant expansion. Contrariwise, between 1914 and 1922, China witnessed an increase in its spindles and looms of over 300 percent, and most of the mills in this period were able to save money and hire Chinese engineers rather than foreign technicians (Chao 1975). Similarly, in 1900 the British‐owned Rio Flour Mills in Brazil reported that through a training program many Brazilians had learned the trade so that “all our millers, engineers, and other skilled workmen, with the exception of less than half a dozen, and all our ordinary workman to the number of about 250, are natives of, or permanently settled in the country” (Graham 1968, p. 139). The Osaka Spinning Company began to produce yarn in 1883 and, “as always, an English engineer came to direct the installation of the spinning machines. A foreign engineer working at the mint in Osaka came to help with the installation of the steam engines. But a Japanese engineer also joined in, so the installation did not completely depend on foreign engineers. The age of complete dependence on foreigners was passing” (Chokki 1979, p. 149).

As machinery development in the North Atlantic became more sophisticated, machinery procurement in “the rest” became more difficult; buying the right machine became an art. Japan's earliest spinning mills “acquired their technical knowledge through foreign engineers who had been sent to Japan to install the equipment and give technical advice.” The Osaka Spinning (p.58) Company, however, set a precedent by sending an engineer to England to learn for himself. When he returned to Japan, he published a book on How to Spin that diffused information throughout the industry. Later, the engineer became the president of the highly profitable Osaka Spinning Company (Chokki 1979, pp. 148–49). In the latter half of the 1880s, all Japan's major cotton mills “became careful about the selection of machines, and sent technical advisors to England in order to investigate various kinds before purchasing” (Takamura 1982, p. 285). The Brazilian proprietor in 1925, by contrast, demonstrated “little knowledge of and less interest in the manufacture of cloth.” Production details were left to foremen “whose sole recommendation was a routine apprenticeship of ten or fifteen years as workmen.” Entrepreneurs “felt that cotton mills were run on the principle of feeding in raw cotton at one end and getting out cloth at the other,” and that “the knottiest problem of a textile enterprise was ‘to order machinery’.” Machinery was ordered by sitting down with a foreign sales representative and discussing the appropriate product to buy (Stein 1955, p. 117; see also Birchal 1999).

Some of the most successful industries, even if minor in overall size, acquired their technology simply through reverse engineering (copying) and studying foreign blueprints. They thrived by catering to small, specialized (often monopolistic) market niches, operating with low inventories and simple cost accounting systems. All this, plus transport costs, made foreign machinery uncompetitive. In Chile, “several large engineering works were in operation in Valparaiso by the end of the century manufacturing [in fact, only assembling] locomotives, railway rolling stock, marine engines, mining machinery, bridges and every other kind of engineering work; the mining machinery was reported to be ‘of a very high class’ ” (Platt 1973, p. 232). In the case of Brazil's light engineering sector in the period 1930–45, which produced customized products with labor‐intensive techniques, some entrepreneurs were Italian immigrants who learned their trade in Brazil while others were Brazilian‐born (the same story applies to Argentina). Whatever their origin, they imbibed an “industrial mentality” at the Institute for Technological Research at the São Paulo Polytechnic Institute. Requirements for administrative and technical manpower were met locally (Leff 1968).

Nevertheless, the Polytechnic in São Paulo graduated mainly civil, not production, engineers. By 1945 it had supplied a total of only four hundred engineers whose degrees were of interest to industry. Industrialists did not lobby the government for more technical education nor did they fund a private technical school because it was cheaper for them to hire foreign technicians (Dean 1969). Technical training everywhere in “the rest” was practically nonexistent.16 Despite its vast mineral wealth, Brazil did not have a school of mines until 1875 (Rippy 1947). In China, an expert study found that the total need for technicians in eighty‐two Chinese mills in the mid‐1930s was estimated to be 4,000 persons, yet only 500 employed by those (p.59) firms had received some professional training in China or abroad. As late as the 1930s, only one college in China was graduating textile engineers (Chao 1975, p. 154). Even in India, the first real institute for technical education, the Victoria Jubilee Technical Institute in Bombay, was not established until the 1880s (with private money). A conference convened in India by Viceroy Curson to expand technical schooling at the turn of the century received a “pathetic” response from local governments and was not pursued further (Kiyokawa 1983, p. 21).17

The neglect of technical education in “the rest” was matched by a neglect of general education. Not surprising, the limited available data indicate that by comparison with the North Atlantic and Japan at a (roughly) comparable period of development, school enrollment, mean years of schooling, and adult literacy rates were much lower in “the rest” (see tables 3.2, 3.3, and 3.4). By 1950 mean years of schooling in “the rest” were not even half of what they were in the North Atlantic in 1913. Adult illiteracy in Argentina and India was higher than what it was in the North Atlantic by a large order of magnitude. All this contrasts with early Meiji Japan, which founded a college

Table 3.2. Mean Years of Schooling—North Atlantic and “The Rest,” 1820–1992

1820

1870

1913

1950

1973

1992

Argentina

4.8

7.0

10.7

Brazil

2.1

3.8

6.4

Chile

5.5

8.0

10.9

Mexico

2.6

5.2

8.2

India

1.4

2.6

5.6

Korea

3.4

6.8

13.6

Taiwan

3.6

7.4

13.8

Average

3.3

5.8

9.9

Belgium

9.8

12.0

15.2

France

7.0

9.6

11.7

16.0

Germany

8.4

10.4

11.6

12.2

Italy

5.5

7.6

11.2

Netherlands

6.4

8.1

10.3

13.3

Sweden

9.5

10.4

14.2

U.K.

2.0

4.4

8.8

10.8

11.7

14.1

Portugal

2.5

4.6

9.1

Spain

5.1

6.3

11.5

U.S.

1.8

3.9

7.9

11.3

14.6

18.0

Average

1.9

4.2

7.7

8.3

10.1

13.5

Notes: Data provided for persons aged 15–64. Blank spaces indicate data were unavailable.

Source: Data adapted from Maddison (1995). Maddison has weighted each level of schooling. Years of primary schooling receive a weight of 1, years of secondary schooling are multiplied by a factor of 1.4, and years of postsecondary schooling are multiplied by a factor of 2.

“The rest” only includes countries for which Maddison provides data. There are no corresponding data for China, Indonesia, Malaysia, Thailand, or Turkey.

(p.60)

Table 3.3. Adult Rates (%) of Illiteracy—Selected Countries, 1850–1990

Country

1850a

1900b

1950c

1970

1980

1990

Germanyd

20

12

na

1

1

1

Sweden

10

na

na

1

1

1

Austrian Empiree

43

23

na

1

1

1

Belgium

48

19

3

1

1

1

U.K.f

32

na

na

3

1

1

France

43

17

4

1

1

1

Italy

78

48

14

6

4

3

Spain

75

56

18

6

7

4

U.S.

20

11

3

1

1

1

Argentina

na

54

14

7

6

5

Brazil

na

na

51

34

26

18

Chile

na

na

20

11

9

7

China

na

na

na

na

35

22

India

na

95

83

66

59

52

Indonesia

na

na

na

43

33

18

Korea

na

na

23

12

6

4

Malaysia

na

na

62

42

30

22

Mexico

na

na

43

26

17

12

Taiwan

na

na

na

na

na

na

Thailand

na

na

48

21

12

7

Turkey

na

na

68

49

34

19

(a) Germany 1849, Austria 1851, Belgium 1856, U.K. 1851, France 1851, Spain 1857, Italy est., U.S. 1870.

(b) Argentina 1895, India 1901, France 1901, Italy 1901, Germany 1871.

(c) Argentina 1947, Chile 1952, Korea 1955, Malaysia 1947, Thailand 1947, France 1946, Italy 1951. Data for Mexico refer to the total population over six years of age. U.S. and Argentina data refer to the population over fourteen years of age. All other data refer to the population over fifteen years of age.

(d) Data for 1850 and 1900 are for Prussia.

(e) Data for 1850 and 1900 are for the Austrian Empire.

(f) Data for 1850 and 1900 are for England and Wales only.

Sources: Data for 1850 and 1900 (for all countries other than India, Argentina, and U.S.) adapted from Cipolla (1969). Data for 1850 measure those who could not read. Both 1850 and 1900 data for these countries refer to total population over ten years of age. Data for Argentina for 1900 adapted from Randall (1977; refers to the population six years of age or older). Data for India for 1850 and 1900 adapted from Lal (1988). Data for the United States for 1850 and 1900 adapted from West (1975, p. 42). This data refer to the total population over ten years of age. Data for 1950 were adapted from UNESCO (1972). Data for 1950 refer to the total population over fifteen years of age. Data for 1970 for Germany, Sweden, Austria, Belgium, U.K., France, and U.S. adapted from World Bank (1976, 1994). All other 1970 data taken from UNESCO (1993). All 1970 data refer to the total population over fifteen years of age (except Malaysia which refers to the population over ten years of age). Data for 1980 for Germany, Sweden, Austria, Belgium, U.K., France, and U.S. adapted from World Bank (1985). All other 1980 data taken from UNESCO (1993). All data refer to the total population over fifteen years of age. Data from 1990 for Germany, Sweden, Austria, Belgium, U.K., France, and the U.S. are adapted from World Bank (1995). All other 1990 data adapted from UNESCO (1993). All 1990 data refer to the total population over fifteen years of age.

(p.61)

Table 3.4. Literate Population in Selected Countries, 1869–1951

Year

Literate Population

Year

Literate Population

(% of Total)

(% of Total)

India

Mexico

1901

5

1910

23

1911

6

1921

29

1921

7

1930

33

1931

10

1943

55

1941

16

1950

57

1951

17

Chile

Argentina

1943

76

1869

22

1895

46

1929

75

China

1943

85

1930

30

Brazil

Turkey

1877

14

1927

10

1942

50

1940

20

1950

33

Notes: Argentina: 1869 and 1895 figures represent percentages of the total population six years of age or older; 1929 figure represents total population over fourteen years of age. Turkey: All data for population six years of age or older. China: Represents male population seven years of age or older. Mexico: 1921 data represents the population five years of age or older; 1910, 1930, and 1950 data represent the population six years of age or older.

Sources: All Indian data adapted from Lal (1988, p. 134). All data for 1942–43 adapted from Hughlett (1946, p. 347). Argentine data for the 19th century adapted from Randall (1977, vol. 2). Data for Turkey adapted from Hale (1981, p. 67). Nineteenth‐century data for Brazil adapted from Graham (1968). Data for China adapted from Rawski (1989, p. 58). All Mexican data (other than 1943) adapted from Wilkie (1970).

of engineering in 1877, vigorously promoted universal primary schooling, and created an elitist university system (students enrolled in higher education rose from 12,000 in 1895 to 127,000 thirty years later) (Kawabe and Daito 1993). Simultaneously, the recruitment and private training of middle managers by big business also improved (Daito 1986).

By way of conclusion, there is virtually no case of a major investment in “the rest's” early industrial history being undertaken without some foreign technology transfer, if only copying. Even the Indian textile industry, with its reputation for indigenous pioneering, started modern operations with a mill in Bombay in 1854 that was partially English‐owned, using techniques and personnel from Lancashire (Mehta 1953). China's first successful chemical works were founded around World War I by Chinese, but the entrepreneurs involved were educated in Tokyo, Kyoto, and Berlin (Rawski 1980). Yet due (p.62) to weaknesses on both the supply and demand sides, technology transfer was a highly imperfect process. This conclusion is based on anecdotal evidence, but no reason suggests such evidence is unrepresentative of the experience of the great majority of “the rest's” technology transfers between 1850 and 1950.

The North Atlantic Contrast

Flaws inherent in technology transfer also afflicted the North Atlantic. “Right up to 1850 and 1860, continental centres frequently failed to achieve British productivity and economy even when using apparently similar equipment” (Pollard 1981, p. 182).18 But the technology acquisition process of the two sets of learners differed. First, because most countries in “the rest” were geographically isolated from the centers of advanced learning, first Great Britain and then other North Atlantic countries, and international transportation and communication were expensive and slow before, say, 1920, “the rest” rarely experienced the mode of technology transfer that North Atlantic countries experienced which were located closer to Britain: not just receiving teachers at home but also sending students abroad. When, for example, Norwegian textile entrepreneurs in the early second half of the nineteenth century began to look to Britain for its supply of equipment, “this led to a series of visits to Britain by virtually all the important Norwegian textile entrepreneurs in search of information and the new technology” (Bruland 1989, p. 61). By 1870 almost one‐third of entrepreneurs in the Rhineland and Westphalia were estimated to have visited Britain on business or to study business (Kocka 1978). The visit of Norwegian and German entrepreneurs was only part of a “grand tour” of foreigners to England: “Prussian, Bavarian, Hanoverian nobles, Russian princes and counts, French marquises, and a medley of Swedes, Danes, Portuguese and Spanish notables pushed their way into Birmingham button factories, swooped elegantly round chemical works, paper mills, munitions foundries or shipyards, and reported their findings back to their ministers at home” (Robinson 1975, p. 3, as cited in Bruland 1989, p. 62).

The disadvantages of learning only by receiving teachers from afar are obvious. One is that learning through an intermediary may be unreliable; firsthand observation is better. Another is that the curious student is denied the opportunity of search activity. Yet there is scant evidence before, say, 1910 that firms in “the rest” sent emissaries to technology suppliers in the North Atlantic. Typical of “the rest” in this period was the way China got its technology for its first steel mill, Han‐yeh‐ping, established in 1896. That the Chinese director of the mill “should have felt fully able to instruct the Chinese Minister in London about the kinds of equipment to be purchased is only an example of the omnipotence that every Confucian official claimed” (Feuerwerker (p.63) 1964, p. 95). The exceptions that support the rule involved unusually well managed companies. To acquire technology for the Tata group's steel mill, Jamshedji Tata, the founder of the group, traveled to the United States in 1907 for four months and conferred with steel experts in Cleveland and Birmingham (he was described by the American press as the “J. P. Morgan of the East Indies”) (Fraser 1919, p. 20). Japan, unlike “the rest”, sent engineers overseas starting with textile manufacturing, as noted earlier. Later, Japan's first steel mill, the Yawata Works, sent a team to a German company for six months (Yonekura 1994). Ironically, this same German company was the technology supplier to Japan's and China's major steel works; the chief engineer in both cases was the incompetent Mr. Toppe. The Japanese ultimately fired him; the Chinese, by contrast, never appear to have learned of his third‐rate skills (Yonekura [1994] and Feuerwerker [1964]).

After World War II, when the costs of learning became cheaper with new and better modes of transportation, communication, and education, technology acquisition by “the rest” increasingly took the form of sending students, workers, engineers, and managers overseas to learn. An excellent example is South Korea's Pohang Iron and Steel Mill (POSCO), which sent hundreds of production and nonproduction workers abroad for hands‐on training (Amsden 1989). Before World War II, comparable training was a serious lapse in “the rest's” education.

Second, the freshness of learning is greatest when it is ongoing. In the case of Norwegian textile firms, “visits to England for the inspection and purchase of equipment were not necessarily confined to the setting‐up period, but continued to be made in order to keep up with technical developments and to purchase further equipment” (Bruland 1989, pp. 65–66). While firms in “the rest” often bought the latest equipment, in some industries capacity utilization was so low that equipment purchases by the same firm were infrequent, especially since bankruptcy rates tended to be very high (see chapter 4). Thus, “it is perhaps an exaggeration to state that the transfer of technology which provided the basis for modern industry in twentieth‐century China was a single‐shot affair, but it does appear that later infusions were slower and smaller than the initial dosage” (Feuerwerker 1967, p. 313).

Third, “the rest's” backward skills and low level of literacy and education made it more difficult—and expensive—to absorb foreign know‐how. The relative expense can be gauged by the length of time a foreign expert stayed at a client's company and the salary she was paid. The aggregate picture for British workers in Norway is one of short stays and high turnover (Bruland 1989). By contrast, the anecdotal picture painted earlier for “the rest” is one of quasi‐permanent foreign managers and long‐duration workers.

Even assuming equal lengths of stay, technical assistants were much more expensive for “the rest” than for the North Atlantic. Spinners in Norway in 1861 earned 34.2 Norwegian shillings while British female spinners in Norway (p.64) earned 49 shillings, or 1.4 more (Bruland 1989). But assuming conservatively that a British worker got twice what she got at home when she worked in “the rest,” then she would earn eight times what an Indian worker earned and ten times what a Chinese worker earned (given that in 1933 the wages per week of a spinner were estimated to be around 71 shillings in the United States, forty shillings in the United Kingdom, twelve shillings in Japan, ten shillings in India, and eight shillings in China, where 1£ = 20 shillings) (Mehta 1953, p. 144). This large wage differential was likely to cause social problems, as it did in the Ottoman Empire, as well as to place a huge financial burden on firms. Moreover, assuming that when locals eventually replaced foreigners in supervisory and managerial positions their salaries were not readjusted downward, then the artificial wage gap between production‐ and nonproduction employees became huge.

Fourth, the scale of efficient operation was rising steadily throughout the nineteenth century. Thus, because “the rest” industrialized later than the North Atlantic, it had to contend with greater problems of scale. Costs of finance, labor relations, and learning, moreover, do not necessarily change linearly with larger scale.

Finally, the list of capabilities that “the rest” was lacking was far longer than that of the North Atlantic. In addition to not having proprietary technology, “the rest” had few established firms. Not only did countries in “the rest” have to import foreign know‐how, they also had to create de novo the organizations to implement that know‐how. Even in Norway, without a history of handicraft textile production, the textile companies that acquired British technology were ongoing companies from one or another background (Bruland 1989).

Beyond Technology Transfer: The Silk Industry

Despite being the inventor of silk and despite a seven‐fold rise in silk output between 1870 and 1928, the Chinese silk industry sustained an extraordinary reversal at the hands of Japan. From 1859–61 China accounted for 50.6 percent of world raw silk exports while Japan accounted for only 6.7 percent (see table 3.5). Between 1927–29 China's share had fallen to 21.9 percent whereas Japan's share had jumped to 67.2 percent (Federico 1994, p. 53). By 1929 Japan's largest single industry was silk reeling, more important than even cotton yarn (Yamazaki 1988).19 Raw silk exports are estimated to have financed no less than 40 percent of Japan's imports of machinery and raw materials between 1870 and 1930.20 Silk manufacture was “a training school for Japanese industrialization” (Hemmi 1970).21 (p.65)

Table 3.5. Export Shares in Silk (%)—Selected Countries, 1859–1938

Country

1859–61

1873–75

1905–7

1911–13

1927–29

1936–38

Italy

26.5

30.9

32.8

19.2

10.3

6.2

China

50.6

53.1

33.9

35.4

21.9

10.7

Japan

6.7

8.3

27.0

41.5

67.2

83.1

Turkey

7.6

4.1

4.8

3.2

0.3

0.0

Other

8.6

3.7

1.5

0.7

0.2

0.0

Source: Adapted from Federico (1994), p. 53.

Both China and Japan modernized their traditional silk industries by importing technology from Italy, including the design of steam filatures. Japan, however, excelled relative to China in raising productivity and quality further. Productivity in Chinese silk manufacture stagnated (measured by the number of reels per basin). It is estimated that China's productivity at the beginning of World War I was lower than Italy's productivity calculated at the time of its technology transfer to China (Federico 1994). The economic success of Japan's silk industry extended beyond adding improvements to imported technologies.22 After studying foreign practices, Japan redefined silk manufacture as a business in response to an industry‐wide crisis.

Silkworm disease (pebrine) appeared in France in 1854 and soon spread throughout the world. It devastated the French silk reeling industry and left France a major importer of raw silk to be woven into fabric. The American silk weaving industry emerged after the Civil War behind a protective tariff never lower than 45 percent and became an even larger importer than France of raw silk (Federico 1994). Both the European and American silk weaving industries preferred silk reeled in steam filatures over silk reeled by hand because the more mechanically spun product was considered superior in regularity, winding, cleanliness, and elasticity. Japan's strength ultimately became that of producing machine‐reeled raw silk of consistent, medium‐grade quality in large, reliable quantities for the fast‐growing American market. By 1909 Japan had overtaken China as the world's leading raw silk exporter despite the fact that at the high end Chinese silk was superior to Japanese silk and despite the fact that in the 1870s and 1880s American buyers had complained that Japanese silk was irregular and inferior (and that Chinese business practices were dishonest) (Li 1981).

One factor behind China's fall was the spread of pebrine disease. Feeble government efforts to eradicate it by forcing peasants to follow a painstaking method devised by Louis Pasteur had failed (Eng 1984). By the 1920s,

it was estimated that the silkworm egg sheets sold on the market were 75 to 95 percent diseased. In Japan and France, one ounce of eggs would yield

(p.66)

110–133 pounds of cocoons, whereas in China one ounce would yield 15–25 pounds. The failure to check this disease was the most critical technological factor in the decline of the Chinese silk industry in the twentieth century. (Li 1981, p. 23)

The erratic supply of eggs raised costs at subsequent stages of production because raw materials were a very high percentage of total costs. At the turn of the century, raw materials accounted for roughly 80 percent of the value of final output (Federico 1994). In 1933 the value of raw material as a percentage of output value was estimated to be higher in Shanghai steam filatures (raw silk reeling) than in any other Shanghai industry (twelve in total) except cotton textiles and flour milling (see table 3.6).

In countries that maintained a supply of healthy silkworm eggs (the Ottoman Empire and Japan), government regulation was instrumental. In theory, private collective action could have substituted for government coordination (Aoki et al. 1997). In practice, government regulation contained an element of coercion that was necessary to force compliance. In the case of Turkey, the residents of Bursa (near Izmir), whose silk manufacture dated to the Byzantine era, banded together and at great cost imported healthy eggs for breeding. Despite initial success, the temptation of high profits led some raisers to engage in “fraudulent practices,” which resulted in diseased eggs and inferior raw silk. The residents' efforts could not overcome “corruption, indifference, and lack of capital” and silk raising at Bursa “appeared doomed.”

Table 3.6. Twelve Leading Industries in Shanghai, 1931–1933

Number of Factories

Average Number of workers per Factory

Value of Raw Material (as % of Output) in 1933

1931

1933

1931

1933

1933

Foundries

35

20

22

36

44.0

Machinery industry

289

160

36

68

44.0

Chemical industry

60

78

128

118

55.0

Matches

3

4

473

402

68.0

Cotton spinning

27

29

2240

2083

75.0

Cotton weaving

61

69

102

129

88.0

Steam ffilatures

66

49

611

607

82.0

Silk weaving

251

115

40

85

63.0

Knitted goods

96

52

80

136

70.0

Rubber goods

29

41

233

269

38.0

Wheat flflour

14

15

160

168

83.0

Cigarettes

44

45

312

388

40.0

Source: Adapted from Lieu (1936, pp. 383‐85).

(p.67) Their knight in shining armor came to the rescue in the form of the imperial Ottoman Public Debt Administration, which was formed in 1881 by European bondholders of the huge Ottoman debt. To raise revenue, the Debt Administration promoted silk production by reviving a silk research station, establishing a model silkworm nursery, upgrading production standards, and—whatever debt holders' ideological commitment to free trade—imposing import duties on silkworm eggs entering Turkey in order to protect local egg growers. Between 1880 and 1908, raw silk exports from Bursa (mainly to France) rose threefold (Quataert 1983, p. 485).23

The role of the government in the silk industry in Japan was at least as comprehensive as that of the Debt Administration in the Ottoman Empire. From the late nineteenth century, it was illegal in Japan for individual households to breed their own eggs; special farmers were licensed to raise eggs and only these could be used in sericulture (Li 1981, p. 24). In addition, as soon as the Meiji government came to power it established inspection offices to ensure quality control and founded research stations to study silkworm diseases (comparable stations appeared in China only in the 1920s, depending on the region) (Hemmi 1970). It has been estimated that between one third and one half of Japan's raw silk growth in the period 1870–1929 was due to the increased availability of inputs (Federico 1994).

The Japanese government also created model silk reeling filatures. These went bankrupt, but technology transfer to the private sector was marked. The Tomioka Filature,24 the first Western‐style silk plant in Japan, was sold to the Mitsui family and then became the Katakura Silk Manufacturing Company, one of the largest and most profitable in the country. The Miyamada model factory inspired a joint management company (Nakayama‐sha) whose technology spread throughout the Suwa region (Togo 1997).

In terms of capital, it was made available to the silk‐reeling industry by government‐controlled financial institutions. With the backing of the Yokohama Specie Bank (est. 1880) and the Bank of Japan (est. 1882), commercial banks were able to assist wholesalers in making loans to silk reelers (Japan's development banks are discussed in chapter 6). By 1917, 94.7 percent of all filatures could afford to use steam rather than charcoal or firewood, up from only 36.7 percent in 1892 (Ono 1986).

Japanese entrepreneurs also managed to overcome organizational problems that had been a source of inefficiency in silk manufacture for generations. Stages in the silk production process, being discrete and decentralized, gave rise historically to a large number of brokers who insured against risk by buffering one production stage from another. But middlemen also myopically shied away from responsibility for the whole process. In China, “those who engaged in the silk business did not necessarily have a long‐term stake in improving the quality of the product. They simply invested their funds hoping to get a quick and high return” (Li 1981, p. 154).25 The nature of (p.68) ownership and control, in particular, contributed to low quality and poor management.26 In the case of filatures, they typically involved partnerships for only one year. Cantonese filaturists rented from the clans, gentry, and landlords who had built “wild chicken” filatures as a form of real estate speculation. The stock company was the typical organizational form, but most investors preferred to spread their capital thin in several ventures rather than concentrate it in one filature, with profits tending to be redistributed rather than reinvested. “The separation of ownership and management considerably lessened the capitalization requirement for operating a filature and made entry into the industry easier, since fixed capital for construction of buildings and machinery had been borne by the leasors or former owners. At the same time, this structural change not only accentuated the lack of horizontal integration of the industry, but also encouraged speculation, promoted instability and discouraged technical renovation and expansion of the plants” (Eng 1984, p. 362). Because of a shortage of credit, rising land values, and escalating construction costs in Shanghai in the late 1920s, 89 percent of filaturists rented their plant and equipment.27

Bigger business units with greater market power evolved in Japan in response to problems of short‐termism, speculation, and decentralized control over the production process. The number of middlemen fell, partially through the formation of rural producer cooperatives but especially through the strengthening of general trading companies (which also controlled China's overseas silk trade). Large, vertically integrated manufacturing companies also emerged. While Japanese filatures generally remained small, after 1910 two vertically integrated companies dominated the machine‐reeling production stage (Katakura in Nagano and Gunze in Kansai), which increased management control over the whole production process and set a benchmark for the rest of the industry. In 1929, in terms of net profits, the Katakura Silk Manufacturing Company ranked thirteenth among Japan's top fifty industrial and service enterprises (Yamazaki 1988).

By way of summary, the Japanese silk industry was the first to introduce a new pattern of production that involved more government and bigger business: government limited entry at the beginning of the production process for purposes of quality control; it financed research; it contributed to organization‐building by investing in model factories; and it created development banks to finance new production equipment. This enabled private enterprise to undertake the three‐pronged investment analyzed more fully in the next chapter. Vertically integrated silk manufacturers emerged in Japan to control the entire output chain, and managerial hierarchies became devoted to motivating labor and monitoring productivity. General trading companies, an extension of Japan's diversified business groups, established the distribution channels necessary to procure raw materials and dispose finished products at highly competitive prices. Technology transfer was thus only the (p.69) beginning of an intensive learning process, even in a traditional, low‐tech industry such as silk.

Conclusion

Every latecomer must learn from an established master. But not all learners are equal. As anecdotal evidence on prewar technology transfer suggests, the more backward the learner, the more difficult the transfer. This tendency perpetuates divergence in income among countries attempting to catch up with the world technological frontier.

Technology transfer to “the rest” before World War II had its moments of mutual gain and cooperation. But it also suffered from weak absorptive skills on the demand side and geographical distance, high costs, incompetence, and ill‐will on the supply side. Foreign investors—first individuals and then firms—typically arrived on the scene after an industry had already been started. They may have raised productivity and quality in the firms they acquired, but usually they did not serve as a catalyst for industrial diversification. No matter how open the economy of a technology buyer, technology transfer proved unreliable as a means to equalize productivity internationally. Ultimately, additional means were necessary—and were mobilized—to narrow the gap.

Notes:

(1.) Tacitness refers to the incomplete specification of technology because (a) its scientific properties are not fully understood so documentation is impossible; (b) its properties are proprietary; or (c) the nature of its properties are more art than science. The first two types of tacitness typically refer to production techniques and hardware—machinery and equipment. The last two types typically refer to software—organizational and managerial capabilties. See Katz (1987) on Latin America, Lall (1987) on India, Mourshed (1999) on the pharmaceutical industry (India and Egypt), Westphal et al. (1985) on Korea, as well as Rosenberg (1982) and David (1997) for a general discussion. Teece (1976) examines the costs that tacitness exacts in technology transfer. For an analysis of the transfer of process technology from an historical perspective, see von Tunzelmann 1997).

(2.) As observed by Cairncross (1962, p. 43, emphasis added), “While foreign investment undoubtedly speeded up the development of (poor) countries, it is more accurate to think of it as accompanying and reinforcing their growth than as preliminary to it. . . . the foreign investor usually did not join in until comparatively late in the day, lagging behind rather than running in front.” U.S. experience “strongly supports” this assessment (Kravis 1972, p. 404).

(3.) Circa the 1930s, textiles and apparel as a share of total manufacturing output were roughly: 26percent in Brazil (Kuznets 1955), 23 percent in Chile (Weaver 1980), 30 percent in Mexico (Bulmer‐Thomas 1994), and 40 percent in China (Rawski 1989). Although Argentina had a textile industry, it was relatively small, around 15 percent of manufacturing output (Weaver 1980). In 1936 spindles in place (thousands) equaled 159 in Argentina, 2,311 in Brazil, and 862 in Mexico. The corresponding figures for looms in place (thousands) were 1.8, 81.9, and 33.2 (International Labour Office 1937, p. 111).

(p.302)

(4.) Their location in “the rest” was influenced by the availability of raw materials. Cotton textiles tended to be produced in countries where raw cotton was grown locally (Japan before the Meiji restoration, Brazil, China, India, Turkey, and Mexico). Steel tended to be made in countries with rich iron ore and coal deposits (they included the above countries except Japan, which sourced its steel inputs from Manchuria, which Japan colonized in the 1930s).

(5.) As noted in the last chapter, the modern Mexican textile industry began in the 1830s, but the first large‐scale mill (the Compañía Industriala de Orizaba, CIDOSA) was founded in 1889 (it was the only mill in 1895 large enough to pay a tax on sales) (Keremitsis 1987). Mills in Brazil's northeast existed in the 1840s, but accelerated development began in other regions in Brazil in the 1890s (Versiani 1980). India's first cotton mill, the Oriental, was established by the Parsi merchant, N. F. Davar in 1854. India's “jewel,” the Tata family's Empress Mills, started to operate with ring spindles in 1877 (Tripathi and Mehta 1990; Tripathi 1982).

(6.) Continental European companies around 1913 had very few manufacturing affiliates in “the rest,” an exception being the Mexican subsidiary of Metallgesellschaft (Germany) (Franko 1974).

(7.) Contrast the early experience of modern sugar refining in Brazil with that in England. In 1875 Henry Tate and Sons joined forces with a German engineer who held a patent for sugar cubes. Previously, grocers had to hack conical loaves into pieces for customers. Tate called on an experienced engineer and within twelve months a refinery had been built. All was not sugar and spice for Tate; he went through a rough financial period and had to withdraw his daughter from boarding school. But he prevailed, and soon the Tate cube became the standard for quotation for refined sugar in London—so great was the confidence in “the lasting nature of the product's quality” (Chalmin 1990, 77).

(8.) Foreign investment lagged rather than led industrialization in imperial Russia as well as in “the rest.” Its contribution was said to have been “decidedly of a minor nature” before 1880. Starting in 1880, it accelerated substantially, and, therefore, it became an important factor earlier than it did in “the rest” (McKay 1974, p. 336).

(9.) See also (Ficker 1995).

(10.) For the minimalist role of foreign investors in China's early industrialization, see Dernberger (1975) and Murphey (1977, p. 126). The latter argues that as late as 1931, “the role played by foreign investment was marginal.” Remer (1933) suggests the role of foreign firms in early manufacturing was substantial but provides little supporting evidence. Hou (1965) provides detailed information on foreign investments in China's early industrialization but does not specify whether domestic firms were also active in the same industries at the time. Feuerwerker (1964) describes early foreign investment in China as “miniscule” in size.

(11.) For Japan, see also Okita and Miki (1967).

(12.) Nanyang was still alive in 1998 despite a highly competitive Chinese domestic market for cigarettes: “The company's major product is well known among Chinese smokers and regarded as one of a dozen high‐end domestic and foreign brand names in China” (Bankers Trust 1998, p. 70).

(p.303)

(13.) In fact, a German company was involved, Gutenhoff‐nungshutte (GHH).

(14.) Clark (1973) as cited in Rawski (1975).

(15.) “All that glittered was not gold” even in the postwar period. In Brazil, a technology transfer by Union Carbide in the late 1960s, for a Wulff naphtha cracker to supply ethylene, resulted in a costly setback: “what had been considered ‘start‐up’ difficulties wherever it had been tried (in countries other than Brazil) turned out to be fundamental flaws. Wulff crackers just did not work” (Evans 1979, p. 233). In India, a chemical explosion in 1984 at a Union Carbide plant in Bhopal resulted in at least 3,000 deaths and 300,000 injuries. The report on the tragedy by the Council of Industrial and Scientific Research attributed it to “failures in design, equipment, supplies, and operating procedures” (Shrivastava 1992, p. 46). In Mexico, the machinery and process technology for a BOF oxygen steel mill purchased from the German firm DEMAG had “a number of problems embodied in the original design, which not only did not correctly take into account the specific conditions of the local environment, but also carried unresolved technical problems of the original design” (Perez and Jose de Jesus Perez y Peniche 1987, p. 187). In Korea, a cement mill was chronically troubled with a technically faulty process supplied by Mitsubishi Heavy Industry. Kolon Nylon invited Chemtex, an American fiber company, to participate in equity and share technology. Production started in 1963 “but confronted many technical difficulties,” whereupon Kolon sought technology from Japan (Tran 1988, p. 399). In cases where foreign firms stumbled, local firms sometimes got the upper hand.

(16.) A failure to invest in education was once held partially responsible for Britain's economic decline, only to be superseded by the belief that a scientific relationship between education and industrial productivity was not empirically established. But “the influence of education may have been too readily dismissed” (Roderick and Stephens 1978, p. 149). On whether education was generally important for growth in light of nineteenth century skill requirements, see Tortella (1990). Kawabe and Daito (1993) analyzes training in modern corporations and business education.

(17.) Technical training was much more advanced in the North Atlantic, and even in Russia: “Despite its relative backwardness, Russia was on the whole fairly well catered for as regards formal technical education even before Emancipation, when its universities and engineering schools were able by 1860 to provide a comprehensive and up‐to‐date training in the main branches of applied science and technology” (Kenwood and Lougheed 1982, p. 109).

(18.) See also Kenwood and Lougheed (1982) and Saul (1972).

(19.) Silk manufacture follows four basic steps: (1) the cultivation of mulberry trees; (2) the raising of silkworms (sericulture); (3) the reeling of silk fiber (raw silk) from cocoons, either by hand or by mechanical steam filatures; and (4) the weaving of silk fabrics. The third step, the focus of competition between China and Japan, is composed of four operations: (a) the drying of the cocoons; (b) the boiling of the cocoons; (c) the reeling of the silk threads from the cocoons; and (d) the re‐reeling of them for finishing.

(p.304)

(20.) In Italy, silk accounted for two‐thirds of exports during the Napoleonic Kingdom (Poni and Mori 1996). In the second half of the nineteenth century, silk was still Italy's biggest export and produced more value added than the chemical, engineering, and metal‐working industries combined (Davis 1991).

(21.) Cited in Li (1981, p. 6).

(22.) Due to Northern Italy's technological capabilities, the Italian silk industry remained a major world producer until it ran out of cheap agricultural labor in the 1910s. Although a latecomer among North Atlantic countries, Italy took the high road in the silk industry by virtue of the innovativeness of its engineering sector—in this it was like the French cotton textile industry analyzed in the last chapter. Italy competed by producing the highest quality silk using advanced equipment. Three innovations were introduced in Italian plants. First, a dryer for silk cocoons was developed between 1885 and 1890 (in response to a contest sponsored by the Italian Ministry of Agriculture, Industry and Trade) which saved labor and capital and raised productivity. Second, a mechanical device to prepare cocoons was invented that saved skilled labor and increased productivity in reeling. Third, a device was developed that mechanically attached the edges of new cocoons to a moving thread, thereby saving time and increasing productivity. From an initial number of two or four reels per basin, by the 1930s the most productive Italian silk factories were handling sixteen to twenty reels per basin (Federico 1994). For some Japanese incremental improvements, see Ono (1986).

(23.) The silk industry was dominated by émigrés from France, Armenia, and other regions to the west. When these émigrés were driven out of Turkey as a result of war and revolution in the early nineteenth century, the silk industry of Bursa collapsed.

(24.) A conservative assessment of the learning contribution of Tomioka is given by McCallion (1989).

(25.) The most well‐known and successful dozen or so silk manufacturers in twentieth‐century China were all concurrently compradors and many also operated their own silk wholesale firms (Eng 1984).

(26.) At the Shanghai filatures, “complaints about the workers' lack of skill and discipline were frequently voiced. One observer said that, compared to the orderliness of Japanese filatures, the Chinese factories were utterly chaotic. Workers were lazy, sloppy, dirty (they combed their hair in the reeling room and boiled corn ears in the cocoon basins), and dishonest” (Li 1981, p. 174).

(27.) The separation of ownership and management in Shanghai's small plants at this time appears to have been general (Lieu 1936).