1

A brief note on regional industrialization in the basic metals sector in China,

1850-present

Bas van Leeuwen, Jieli Li, Zipeng Zhang

1. Introduction

In this brief note we report our progress on deriving estimates of the basic metal sector, i.e. the smelting

of ferrous (iron &steel) and nonferrous (other metals than iron &steel). We would like to stress this is

progress report rather than a finished (or even unfinished) paper.

Our choice for the basic metal industry is based on its importance for economic development.

Indeed, the better availability, higher quality, and lower price, aided in the development of initially

boilers, railways, steam engines and, subsequently, a wide variety of machine tools aiding economic

development. Indeed, this importance for economic development was also recognized by the Chinese

government during the movement of Great Leap Forward (1958-1961) aimed at rapidly industrializing

China. During this period, steel production was considered as one of the main indicators of

modernization and pig iron production rose from ca. 6 million tons to 27 million tons. Although

subsequently a collapse occurred, it was followed by fierce growth. China’s crude steel production

accounted only 2.8% (International Iron and Steel Institute, 1978: 3) of the global total in 1967. As of

today, this share has risen to almost 50% (The World Steel Association, 2017: 2).

In the following Sections we deal with definition and sources (section 2), data (section 3), some

preliminary results (section 4) followed by a brief conclusion.

2. Definition

A first question that may come to mind is how to define the basic metal industry. We do this by using

the International Standard Industrial Classification (ISIC) revision 4 from the United Nations (2008). This

classification, which can be traced back to its inception in 1948, is the main vehicle for comparing

industrial data over time as most countries either follow ISIC definitions, or use a classification scheme

that is based on ISIC. Within revision 4 we can find the basic metal industry under Section C

(manufacturing), Division 24 (Manufacture of Basic Metals) where it is defined in Group 241 and Group

2

242 as the “smelting and/or refining ferrous and non-ferrous metals from ore, pig or scrap”. This, in

turn, may be modified in alloys by adding chemicals.

It is important to note that above definition only includes the smelting and refining of both

ferrous (Group 241) and non-ferrous (Group 242) metals. It therefore excludes Group 243 (the casting of

metals), which is commonly included in the metallurgical sector. The main reason for excluding casting is

that in China there is no clear distinction in the casting of intermediate goods like pipes and wires

(Group 243) and more refined metal casting (division 25 in the ISIC) as, contrary to many other

countries, China does not use the ISIC classification or a derivative (see e.g. Holz 2013).

Looking at the smelting and refining part of basic metals only, as pointed out by Holz and Lin

(2001), from 1984 onwards this was aggregated in the Smelting and rolling of ferrous metals 1 and

Smelting and rolling of non-ferrous metals 2. Hence, for our study we have to remove the “rolling” from

both sub-sectors. Before 1984, even though little information exist, Holz found that most classifications

included smelting only in aggregated form as Metallurgical industry3, i.e. smelting, rolling, and casting.

Even more complicated are the pre-New China periods, i.e. Republic of China (1912-1949) and Qing

China (1644-1911). For those periods, the manufacturing of base metals was not classified in

manufacturing, but rather in mining. Therefore, we had to go to every individual producing region or

even enterprise, and try to distill the information of smelting of ferrous and nonferrous metals.

3. Data

It is important to note that the sources are quite different for our three benchmark years, i.e 2004,

1933, and 1850. For 2004, we used the 2004 economic census which contains for both above and below

“scale” (5 million yuan revenue) factories the main product, number of laborers, and revenue. It is

important to stress that this excludes the self-employed. Yet, comparison with data for e.g. Nanyang

prefecture in 2004 ferrous metal shows that the share of self-employed in metallurgy is only 1.5% of

revenue and number of laborers although the share of factories is, in some cases, substantial (Nanyang

Municipal Bureau of Statistics, 2006). But we focus on smelting, instead of total ferrous metal which

includes casting, where the share of self-employed is even lower. The total number of laborers and

revenue by product and region is calculated based on the main output product of each company. That

1 黑色金属冶炼及压延加工业

2 有色金属冶炼及压延加工业

3 冶金工业

3

is, a factory that has a main product “steel” is assumed to only produce steel. This is a big assumption, of

course, but since this applies to all regions, the effect cancels out to some extent as is also shown by

some preliminary comparison with prefectural yearbooks. For total output in tons of metals, we used

the estimates from the various prefectural, provincial yearbooks. There are two problems. First, these

output figures are mostly above scale only, but since we know revenue both below and above scale, we

used those to estimate the missing part, mostly being not more than 5% of total output.

Compared to the post-1949 era, regional industrial data, especially for administrative units

below province, of the two earlier benchmarks are either far less directly available or even absent in the

sources. This forces us to practice a historical GDP-style approach to construct the datasets. For 1933

mercury, silver and mechanical production of pig iron are taken from the General statement on the

mining industry of China edited by Hou (1935). Most native handicraft production of pig iron is

estimated based on the same source. The exception is Zhejiang, which is estimated based on the

Ministry of Industry (1933) and Hou (1935). Copper and steel are estimated based on Hou (1935). For

antimony, Guangxi is estimated based on Guangxi Mining Bureau (1935), and Hunan is taken from Hou

(1935). For tin, Guangxi is estimated based on Guangxi Provincial Bureau of Statistics (1936), Hunan,

Jiangxi and Yunnan are taken from Hou (1935), and Guangdong is estimated based on Hou (1935). For

lead and zinc, Yunnan is estimated with Hou (1935), Hunan is taken from a contemporary survey (1934),

and Sichuan is estimated based on government archives (The Second Historical Archives of China, 1994).

For gold, Gansu and Qinghai are estimated based on survey (1935), Guangxi is proxied with 1932 data

from Guangxi Provincial Bureau of Statistics (1934), Guangdong, Hebei and Rehe are estimated with Hou

(1935;1932), Henan is taken from a survey (Wang, 1934), Shandong, Sichuan, Heilongjiang, Jilin and

Liaoning are estimated based on Hou (1935), Hunan is taken from the Ministry of Industry (1935),

Xinjiang is estimated based on Xu (1944). In order to get total output value and number of laborers, we

require prices by product as well as output/worker by product. The prices are taken from Ou (1947). The

productivities of pig iron (both handicraft production and modern mechanical production), copper, gold,

lead, steel, tin, antimony and zinc, are taken from Ou (1947), and silver is proxied by gold. For mercury,

production is calculated with output and worker number taken from Ou (1947).

In general we based our 1850 data on the Qing central government archives, e.g. memorials to

the throne, which are accessed from the First Historical Archives of China in Beijing, the Institute of

History and Philology (Academia Sinica) in Taipei, and the Palace Museum in Taipei. These reports to the

emperor, in certain cases provide information about factory, product, region, and sometimes output

4

and labour. But most often, the output is not directly recorded and hence has to be estimated with

other given information, e.g. taxation. It is very much work in progress, hence we also base part of our

estimates on reworking of existing studies. The production in tons are re-estimated based on Xu and

Zhang (2015), whose results were based on various and massive sources, including memorials to the

throne, government books, local gazetteers and surveys, but were presented in provincial level. The

prices are taken from Xu and Zhang (2015) as well.

Productivity/worker in 1850 is more difficult. For copper, it’s calculated with Peng (1985)'s

estimation on copper for the Qing dynasty. The result for copper is used as the proxy for cupronickel.

For mercury, it’s proxied by the calculated 1933 productivity. Tin is proxied by Ou (1947)'s estimation of

Yunnan (without modern producing) of 1934. Lead and zinc are proxied by Ou (1947)’s estimation for

1933. Gold is proxied by Ou (1947)’s estimation for 1933 and converted from tael (500g) to Qing tael

(596.8g). Then we use the result of gold to estimate silver. For steel, it’s calculated based on Xu and

Zhang (2015) and An (1941). The productivity of pig iron in 1850 is even more complex than the other

products, as regional difference was so significant that we had to apply specific rates for each province.

For Sichuan, Shaanxi and Fujian, the productivities are taken from Xu and Zhang (2015). Guizhou and

Hunan are calculated based on Xu and Zhang (2015). For Shanxi, it’s calculated based on Xu and Zhang

(2015) and Fan (1985). Guangdong and Guangxi are calculated based on Xu and Zhang (2015) and Qu

(1687). The productivity of Jiangxi is proxied by Zhejiang which is calculated based on Cao and Jiang

(2010). Yunnan and Anhui are proxied by Ou (1947)’s estimation for 1933.

4. Geography

Our data are in principle on county level in contemporary boundaries. However, given the errors in the

data, we thought it referable to only report the prefectural data, i.e. about 300 regions. Needless to say

boundaries changed strongly between 1850 and present, if only because Outer Mongolia became an

independent country in the first half of the 20th century. We therefore used the 1820 prefectural map

from CHGIS for the benchmark 1850, but the 1930s and 2000s map were created within our project and

consist, besides from political boundaries, also from roads, railways, waterways and mountains.

Our aim is to create concordance tables that allow creating more or less consistent boundary

estimates over time, but that will take about another month to complete.

5

5. Changing patterns

Please note that we are in the process of upgrading. Amongst others:

a) Some Northern provinces are lacking in 2004;

b) Some map-related prefectural distributions in 1933.

That being said, figures 1-3 show TENTATIVELY that there was some shift to the East of China

between 1850 and present. The main reason was

a) A rise in ferrous metal (Iron and steel) that was located mostly in the East and north.

b) A move towards the centers of railways and waterways (for ease of transport).

c) Other studies also show that more human capital existed in the East.

Figure 1. Number of metal smelting workers by prefecture 1850

Figure 2a. Number of metal smelting workers by prefecture 1933

6

Figure 2b. Number of metal smelting workers by prefecture 1933 & railways & waterways

Figure 3a. Number of metal smelting workers by prefecture 2004

7

Figure 3b. Number of metal smelting workers by prefecture 2004 & railways & waterways

Figure 4. Output of iron & steel in 1850

8

Figure 5a. Output of iron & steel & railroads and waterways in 1933

Figure 5b. Output of iron & steel & railroads in 2004

9

6. Conclusion

This was a brief description about our ongoing work. So every conclusion drawn from the data is

tentative.

For each of the three benchmark years we use different sources ranging from (mainly but not

only) memorials to the throne (1850), to scattered sources in an emerging statistical environment (1933)

and the economic census (2004).

The main preliminary conclusion seems, besides a strong rise in especially ferrous smelting, a

move to the East and North.

10

Reference [1]"the yield table of Shuikoushan in 1933." Miner's Weekly, no.307, 1934: 1061-1062. [2]"production and market price of gold in Gansu." Miners' Weekly, no.319, 1935: 103. [3]Holz, Carsten A. and Yi-min Lin. "Pitfalls of China's industrial statistics: inconsistencies and specification

problems." China Review, 2001: 29-71. [4]International Iron and Steel Institute, ed. A handbook of world steel statistics 1978, Brussels: International Iron

and Steel Institute, 1978. [5]The World Steel Association, ed. Steel Statistical Yearbook 2017, Brussels: The World Steel Association, 2017. [6] United Nations. International Standard Industrial Classification of All Economic Activities (Revision 4). New

York: United Nations, 2008. [7]An, Chaojun. "native steelmaking in Sichuan." Report of Iron and Steel, no.1, 1941: 127-130. [8]Cao, Shuji and Qin Jiang. "Southern Zhejiang Rural Industry and Markets during the Qing Dynasty: Evidence Derived from the Iron Smelting Industry in Shicang Village." Bulletin of the Institute of History

and Philology Academia Sinica, vol.81, no.4, 2010: 833-888. [9]Fan, Baisheng. "Investigation report of crucible iron making in Jincheng, Shanxi." In Essays on the history of

technology, vol.13, edited by Institute for History of Natural Sciences, Chinese Academy of Sciences. Shanghai:

Shanghai Science and Technology Press, 1985, pp. 143-149. [10]Guangxi Mining Bureau. "the output of antimony Regulus in Fu-He-Zhong of Guangxi in the recent three years."

Miner's Weekly, no.362, 1935. [11]Guangxi Provincial Bureau of Statistics, ed. Guangxi Yearbook, Issue 2, Taipei: Wenhai Press Company, 1936. [12]Guangxi Provincial Bureau of Statistics, ed. Guangxi Yearbook, Issue 1 "Guangxi Provincial Bureau of Statistics",

1934. [13]Hou, Defeng, ed. General statement on the mining industry of China, Issue 5, Beijing: The Geological Survey of

China, Ministry of Industry, 1935. [14]Hou, Defeng, ed. General statement on the mining industry of China, Issue 4, Beijing: The Geological Survey of

China, Ministry of Industry, 1932. [15]Nanyang Municipal Bureau of Statistics, ed. Nanyang Economic Statistical Yearbook 2006, Beijing: China

Statistics Press, 2006. [16]Peng, Zeyi. "The labor productivity of mining and smelting in the copper, lead and zinc mines of the Ming and

Qing Dynasties." In Pingzhun Xuekan, Issue 1, edited by SUN, Y.. Beijing: China Commerce Publishing House, 1985,

pp. 372-382. [17]Qu, Dajun, Guangdong Xinyu, vol.15, 1687. [18]Ministry of Industry, ed. The industrial gazetteer of China, Zhejiang Province Ministry of Industry, 1933. [19]Ministry of Industry, ed. The industrial gazetteer of China, Hunan Province, 1935. [20]Wang, Jingzun, ed. Henan mining report, Issue 1: 1933, Kaifeng: Henan Provincial Geological Survey, 1934. [21]Ou, Pao-san, ed. National Income of China, 1933, Part 1, Shanghai: Zhonghua Book Company, 1947. [22]Ou, Pao-san, ed. National Income of China, 1933, Part 2, Shanghai: Zhonghua Book Company, 1947. [23]Xu, Yi and Zipeng Zhang. "A Valuation Study on China's Mining Industry in the Middle of the 19th Century." In

Qing History Forum, edited by Qing History Research Group, Institute of History, Chinese Academy of Social

Sciences. Beijing: Social Science Academic Press, 2015, pp. 165-210. [24] Xu, Chonghao. Xinjiang zhi lue. Nanjing: zhengzhong shuju, 1944. [25]The Second Historical Archives of China, ed. Republican China, A Collection of Archival Documents, Volume

5, Book 1, Part 5: Industry and Mining, Nanjing: Jiangsu Ancient Books Publishing House, 1994.