Jump to ContentJump to Main Navigation
The Industrialization of Rural China$

Chris Bramall

Print publication date: 2006

Print ISBN-13: 9780199275939

Published to Oxford Scholarship Online: January 2007

DOI: 10.1093/acprof:oso/9780199275939.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2017. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see http://www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 26 February 2017

(p.344) APPENDIX 4 The Definition of Rural Industry

(p.344) APPENDIX 4 The Definition of Rural Industry

The Industrialization of Rural China
Oxford University Press

The classification of industry in China in the late 1970s was made more straightforward than in other countries by the virtual absence of private and foreign-owned enterprises. This left two main categories. State-owned enterprises (SOEs) were owned by higher level administrative jurisdiction—central ministries, provinces and provincial-level cities, prefectures and prefectural-level cities, counties and county-level cities—and they were ‘owned’ in the sense that residual profits (after any agreed retentions for investment and wages, and after the payment of taxes) accrued to the owning jurisdiction. Collectively-owned enterprises (COEs) were enterprises where residual profits notionally accrued to the workers and managers of the enterprise, instead of the owning level of government. The category thus covered central ministry-owned, city-owned and county-owned collectives (jiti gongye qiye), enterprises owned by communes (gongshe) and brigades (dadui), usually abbreviated as shedui enterprises (CBEs), and enterprises owned by towns located within counties (zhen).1

This distinction between the state and collective enterprises was misleading in two respects. Firstly, most urban collective enterprises were little different to SOEs in scale, level of technology and way in which they were administered.2 They were, for example, often fully incorporated into the Plan, and any residual profits were remitted upwards to the owning level of government. Indeed some collective enterprises were more tightly controlled by upper levels of government than the SOEs operating alongside them (Blecher and Shue 1996). Secondly, CBEs were collective in name only. As communes and brigades were organs of government, it is more appropriate to view CBEs as state enterprises also. The distinction between them and other SOEs can best be thought of in terms of the level of government which owned them; CBEs were owned by local (i.e. sub-county) government, whereas other enterprises were owned by higher level jurisdictions.

(p.345) A4.1. The Re-definition of the Commune Sector

The main post-1978 change in classification was caused by the re-organization of local government in March 1984 (set out in the Central Committee's document no. 4 of 1 March 1984), whereby xiang (townships) and zhen (xiang-level towns) replaced communes, and cun (villages) replaced production brigades; production teams (dui) retained their name but lost their political and administrative functions.3 As a result, the ownership of CBEs was transferred to the new xiang and zhen; the industries thus remained (at least initially) under the ownership of the local state.

However, the new category of xiangzhen qiye (literally township and town, but usually translated as township and village) was broader than the old CBE category. It included smaller cun yi xia bian qiye (sub-village enterprises) as well as the old commune and brigade industries; these smaller enterprises were owned by dui (teams), lianying (cooperatives) and geti (individual or self-employed households). These three enterprise types had been included in the sideline sub-sector of agriculture before 1984—that is, they contributed to agricultural rather than to industrial production in the Chinese national accounts. The 1984 change included them in the TVE category, and thus re-classified their output as industrial instead of agricultural. The coverage of TVE sector was further expanded to include private (siying) enterprises after their legalization in 1988.4 In addition, industrial TVEs included industry owned by zhen because many of the zhen were seen as little different to xiang.Zhen-owned industry is often therefore often included in the xiang category in Chinese statistical publications. The term xiangcun industry also appears in Chinese materials. In general this differs from TVE because it excludes below-cun industry and zhen industry, and it is often used to refer to the public component of the TVE sector.

Before 1984, we are justified in regarding CBEs as enterprises owned by local government. From 1984 to about 1996, however, we should think of the successor TVEs as being subordinated to local government. Private enterprises were by definition not owned by the xiang,zhen or cun governments. Further, as a number of writers have pointed out, some enterprises used the term ‘collective’ as a flag of convenience to describe themselves and thus have ‘red hat’ status. They were de facto private enterprises but wished to insure (p.346)

Table A4.1 The township and village sector in the mid-1980s (number of enterprises; millions)





Township (xiang/zhen/she)





Village (cun/dadui)





Team (dui)





Cooperative (lianhu)





Self-employed (geti)










Note: These data include all types of TVEs; 240,000 of them were involved in agricultural activities in 1986. The geti figure is a residual. The township figure here includes towns.

Source: Nongye bu (1989: 290–1).

themselves against both a change in the political wind, and continued hostility towards the private sector within the Chinese Communist Party. In addition, collective status exempted them from the taxes levied on the private sector. After 1996, TVEs comprised a dwindling number of public enterprises because of the rapid privatization of xian,zhen and cun enterprises; most enterprises included in the TVE category after 2000 were certainly private. Nevertheless, we are probably justified in using the phrase ‘subordinated to’ local government for these enterprises during the bulk of the 1980s and 1990s. Even the private sector was subject to what might be called indicative planning during these two decades; it was a brave entrepreneur indeed who ignored the wishes of local government given the extra-juridical power it was able to wield in respect of finance, taxation and the regulation of the labour force.

The size of the industrial TVE sector by the mid-1980s is set out in Table A4.1. It is clear that most of the dramatic increase of the mid-1980s was driven by re-definition, namely the inclusion of small self-employed household businesses from 1984 onwards. Not surprisingly, the contribution of xiang and cun enterprises to total TVE output and employment was much larger than suggested by the number of enterprises. In 1986, in fact, xiang and cun enterprises together employed 55 percent of all TVE workers and contributed 69 percent of gross output value, even though they accounted for only 10 percent of enterprises (Nongye bu 1989: 290–5).

A4.2. The Limitations of the Shedui and Xiangzhen Categories

It is conventional to define the rural industrial sector of the late Maoist era (1958–78) as encompassing industrial commune and brigade enterprises (CBEs), and that of the post-Mao era as encompassing township and village industrial enterprises (TVEs). However, there are five problems with this approach which I will outline in turn. Two are relatively minor: the omission of household sidelines, and the inclusion of agricultural enterprises in the TVE category. The third and more serious problem is that it is often unclear whether CBE includes (or is even meant to include) industries owned by production teams. Fourthly, the CBE category excludes industries owned by town (zhen) governments. These last two criticisms do not apply to the xiangzhen concept (which includes zhen and all types of small-scale industries except sidelines), but is a weakness in the CBE approach. The fifth problem—and by far the most serious—is that both TVE and CBE categories (p.347) exclude industries owned by the county (and higher levels of government) but based within county boundaries.

A4.2.1. Household Handicraft Sidelines

Even the TVE definition excludes household handicraft sidelines (nongmin jiating jianying shougongye). Insofar as household handicrafts were sideline products rather than the main products of households, they continued to be classified as sidelines in 1984, and continued to be so treated after the sideline subsector was itself incorporated into farming in 1993. They have therefore been consistently classified as a component of agricultural output in the Chinese accounts, even though it is arguable that they ought to be classified as a type of rural industry. Of course one might argue on the contrary that, precisely because such handicrafts are sideline products of households engaged primarily in agricultural production (and often closely related to agricultural products), they should be classified as part of agricultural production.5 But none of these issues are especially worrying because the subsector is simply not large enough to distort trends. Although the absolute value of rural handicraft sidelines was large—17.2 billion yuan (1980 prices) in 1986 (Nongye bu 1989: 122), and 36.5 billion yuan (1990 prices) in 1992 (NCTJNJ 1993: 67)—these sidelines contributed only 4.2 percent of net peasant income in 1978, 3.1 percent in 1988 and 2.7 percent in 1999 (Xin nongye 2000: 75, 85). It is hard to get excited over these sorts of magnitudes.

A4.2.2. Agricultural Enterprises

The second problem with the TVE category as a measure of rural industry is that it includes some farm enterprises. If we wish to identify the number of TVE industrial enterprises (xiangzhen gongye qiye), we need to exclude farm TVEs. In 1978, agricultural enterprises made up around one third of all xiang and cun enterprises and employed 22 percent of the xiangcun workforce. Thereafter, their decline was rapid in both absolute and relative terms; agricultural xiangzhen employment was only about a third of its 1978 level by 1986, and its share in the employment total was 3 percent (Nongye bu 1989: 290–2). Their inclusion in the TVE category matters even less by the end of the 1990s; only 165,000 of the 20.7 million TVEs were agricultural in 1999 and they contributed only 1.4 percent of value-added (NYNJ 2000: 246, 250). For the late 1970s and 1980s, however, we need to strip out farm enterprises from the CBE total to track the development of rural industry. Fortunately, these sorts of disaggregated data are usually available.

A4.2.3. Sub-brigade Industry

The third, and more serious, problem relates to the coverage of CBE industry. Some accounts use CBE to include team (dui), cooperative (lianhe), individual (geti) enterprises and even siying (private) enterprises. The expanded use of CBE in this way in effect applies the coverage of the term TVE to the industries of the Maoist era and thus ensures a degree of consistency of treatment.

However, the data which appear in most publications and which are labelled shedui usually do not include these small enterprises. In fact, there are no consistent national data (p.348) on non-agricultural enterprises operating below the brigade level before 1984—see for example Nongye bu (1989: 291) or He (2004: 39). Some Chinese jurisdictions have revised the late Maoist data, re-labelling commune and brigade industry as township and village industry, and expanding the coverage of the series to include team-level industry and below. However, this is probably true of only a minority. Brigade-level production is often omitted from the CBE category in most County Record data, and small-scale enterprises are often just ignored. In part this is because the data do not exist; these were usually regarded as little more than agricultural sidelines. However, this type of omission also reflects a continuing reluctance to acknowledge the existence of private industry in the Maoist era.

Irrespective of the coverage of the CBE category in Chinese sources, any plausible definition of rural industry must include team, cooperative, individual and private industry. For example, private industries existed in rural China during the early 1960s, and re-emerged in the mid-1970s in provinces such as Zhejiang, Guangdong and Anhui; it would be wrong to ignore their existence, or that of the more numerous industries owned by production teams. As far as possible, therefore, the usage of the term ‘rural industry’ in this book covers all types of sub-brigade industry.

A4.2.4. Towns

A fourth definitional problem relates to industry operating within towns (zhen). Here there are two separate questions. Firstly, should this type of industry be classified as rural? Secondly, was zhen industry included in the CBE aggregates?

Consider first the treatment and designation of towns in the 1980s and 1990s. One use of zhen is to describe relatively small towns located within xiang. These are usually the seat of the xiang government, just as they were the seat of commune government in the late Maoist era. Such towns are often called jizhen (market towns) and are clearly sub-xiang economic entities. In that sense, their industrial production clearly features within xiang aggregates and therefore causes few problems. However, there were other zhen within Chinese counties that had genuine administrative status, and therefore the term TVE industry has been used after 1984 to cover industries owned by designated towns (jianzhi zhen or simply zhen), county towns (xianzhen) and townships (xiang). These first two need some explanation.

Firstly, consider county towns. They existed because the seat of government in almost all Chinese counties was designated a zhen. Such towns have a variety of names across China e.g. xiao chengzhen (small city), xiancheng (county city) or xianzhen (county town); all owned industries of various types. Some of these industries were owned directly by the zhen. Others were owned by jiedao (neighbourhood or district) committees in instances where the town was comparatively large. Jiedao industries in some ways therefore correspond to brigade enterprises in township or communes.

The second category is that of designated towns (jianzhi zhen). In many poor counties, especially during the early 1980s, the county town was often the only jurisdiction bearing the zhen appellation. Indeed Fei (1986: 103) noted that this was true even for many of the counties of northern Jiangsu. However, it was certainly not true of more prosperous Chinese counties, many of which were home to a number of towns that were approximately equivalent in status to communes; these zhen are often known as jianzhi zhen (designated town). After the 1984 re-organization of local government, which saw the communes renamed as xiang, industries located in these zhen—as well as industries in the county town—were included in the xiangzhen category. More precisely, according to (p.349) the set of rules published in November 1984, a xiang jurisdiction was up-graded to zhen status if it was the seat to the county government, or if it was the seat for the government of a xiang (proved that the xiang had a population of less than 20,000 and more than 2,000 non-agricultural persons resident in it), or if it had a population of more than 20,000 and a non-agricultural population of more than 10 percent), or if it were special in some other way (e.g. a mining area) (Gong'an bu 1985: 102). In the case of Jiangsu province's Suzhou municipality, there were 17 zhen scattered across the five counties which made up the municipality (Shazhou, Taicang, Kunshan, Wuxian and Wujiang) (Gong'an bu 1985: 27). By 1995, the rules had changed substantially. As a result, all five of Suzhou's counties had become cities (shi) and every one of their xiang had been abolished or upgraded, leaving a total of 152 zhen (Suzhou tongjiju 1996: 12). The total non-agricultural population certainly increased over this period but not at a rate sufficient to justify the re-classification of xiang as zhen. In other words, in Suzhou, and elsewhere in China, most of the zhen which existed by the mid-1990s were not true urban jurisdictions but essentially xiang-level rural units.

For the post-1983 era, therefore, we can be confident that the xiangzhen category included industry based in all types of jurisdictions called zhen, irrespective of whether they were xianzhen or jianzhi zhen. A case could be made for categorizing xianzhen industry as urban, but even county seats—especially in the early 1980s—had a profoundly rural character to them. One thinks here of the absence of pavements, and frequently unmetalled roads. As for the jianzhi zhen, most of them were upgraded to zhen status in the 1980s and 1990s simply by administrative fiat. Until the mid-1990s at least, few deserved the name. For these reasons, it is clear that the industries classified as xiangzhen genuinely deserve to be called rural.

What of the Maoist era? Some accounts steer around this problem by simply assuming that the only jurisdiction which existed immediately under the county in the late Maoist period was the commune. This no doubt partially reflects the tendency in Chinese for the term zhen to be pressed into service in a variety of contexts (see for example Mo 1987: 16), and the terminology here is very confusing. Nevertheless, it is very clear that a considerable number of towns existed as administrative units directly under counties in the late Maoist era. In Kunshan county (Jiangsu), there were 21 communes and three towns in 1963 (Kunshan XZ 1990: 87). Pengxian in Sichuan was home to seven zhen and 30 xiang (later communes) in 1963, and to six zhen and 32 communes in 1981 (Peng XZ 1989: 31). Much poorer Wushan county in the Yangzi Gorge region was divided into seven qu (districts) in 1963 and the one zhen and 64 communes were allocated to one or other of the qu (Wushan XZ 1991: 41). The presence of a single town in Wushan thus replicated the type of pattern found in northern Jiangsu, and the qu also appears as an intermediate level in other counties; for example, Zhejiang's Linhai county was divided into districts and was home only to two zhen even in 1982 (Linhai XZ 1989: 90). Ningxian (Gansu) seems to have been an even more extreme case. It had no administrative towns, simply 22 communes, in 1978 (Ning XZ 1988: 54).

The industrial production of enterprises owned by these towns played an important role in the local economy in late Maoist China, and cannot simply be ignored. We can glean an impression of its significance from those County Records which have explicitly distinguish between zhen and commune enterprises. Of the 482 brigade-level and above industrial enterprises operating in 1980 in Gansu's Tianshui municipality—the city proper plus the counties under its jurisdiction—208 were xiang enterprises and 41 were zhen enterprises (Tianshui tongjiju 1989: 524). A fuller accounting is provided for Taicang (p.350)

Table A4.2 The structure of county industrial production in 1978 (percentages of county GVIO)






County collectives












Source: Taicang (1991: 270); Taixing XZ (1993: 266).

and Taixing counties in Jiangsu (Table A4.2). These data show that in both cases the zhen subsector accounted for nearly 10 percent of all industrial production in 1978. Jiangsu was probably unusual in that it had more town-owned enterprises than most other provinces, but it is nevertheless clear that the zhen sector cannot be ignored in any proper accounting.

The significance of zhen industry means that it is galling that the coverage of the CBE category is so imprecise for most counties in the Maoist era: the data published for Taicang and Taixing are unusual. However, it is likely that the CBE category often omits zhen industry, thereby ignoring industry which was both important and rural in character. CBE industry is therefore too narrow a definition of Maoist rural industry.

A4.2.5 Re-thinking the Status of County SOEs and COEs

The most serious problem with both CBE and TVE definitions of rural industry is that they omit SOEs and COEs located within county-level jurisdictions, whether xian (counties) or county-level cities (xianji shi).6 Most of these industrial enterprises were located in either the county town or in another town within the county (few were located in communes) during the late 1970s.

Some scholars have argued that the exclusion of such a large number of essentially rural enterprises from the definition of rural industry distorts underlying Chinese realities. Mo (1987: 16), in his well-known study of TVEs in Jiangsu, was explicit in contending that rural industry was a broader entity: ‘Xiangzhen industry is only a component part of rural industry’. Christine Wong (1991a: 321), in arguing that much supposedly self-sufficient rural industrialization during the Cultural Revolution was actually state-funded, includes agriculturally-orientated SOEs at county and prefectural level in her definition: ‘In this chapter “rural industry” will include industrial enterprises run by people's communes and production brigades, as well as state-owned industries at the county and prefectural levels that are oriented toward agriculture’. She delineated the rural industrial sector even more broadly in an earlier paper, her justification for the inclusion of local SOEs and COEs in the rural category being that ‘These enterprises are mostly small-scale and operate largely free of state control’ (Wong 1989: 39). On the basis of this approach, the rural sector contributed 25 percent of national GVIO in 1983; large and small collectives owned (p.351) by counties and prefectures contributed 3–5 percentage points, state-owned enterprises at the county level contributed 13–15 and CBEs a further 7 points (Wong 1989: 38–41). Thus the CBE sector made up only about a third of the rural industrial sector. Riskin (1978a: 77) also preferred a broader definition: ‘“Rural industries” refers to those industries run by the rural communes and their production brigades, and state industries at the sub-provincial level, chiefly those operated by the hsien’.7

The essential problem with the official classification in the eyes of Mo, Riskin and Wong is that many of the county-owned SOEs and COEs operating in county towns and other larger towns within China's counties had more of an affinity with the rural than the urban sector. The county towns of the 1980s had little in common with China's big cities in terms of infrastructure or facilities. We can thus apply the logic of the previous section. Zhen enterprises fall within the TVE definition of rural industry. Given that most SOEs and COEs were also located within these towns, it makes sense to classify these too as rural enterprises. Furthermore, local-level COEs and SOEs were integrated into the rural economy in three respects. Firstly, a considerable part of enterprise revenue entered the county budget and was therefore available to finance a range of rural projects. Secondly, much of the expenditure by enterprise employees went on goods sold in local markets and produced by the local farm population. Thirdly, a proportion of the SOE and COE workforce in the 1970s and 1980s was made up of ‘temporary’ contract workers (linshi hetong gong) whose hukou was agricultural and who in fact were not temporary but long stay workers.8 These workers were recruited from the communes to fill labour shortages in SOEs and COEs; they were attractive to these enterprises because their hukou status was in the countryside, and thus the enterprises did not have to provide the housing and educational facilities which were the norm for contracted SOE and COE workers.

If we accept that county-owned SOEs and COEs are rural, a logical extension of the argument is that SOEs and COEs owned by higher level jurisdictions (central ministries, provinces, prefectures and municipalities) but located within counties should also be classified as rural. These enterprises were by no means unimportant, especially in the 1980s before the explosive growth of the xiangzhen sector.9 Moreover, there is a case for (p.352) arguing that they should be classified as rural enterprises on the grounds that their workers contributed to the local economy on the demand side, and that these enterprises also recruited temporary workers from the agricultural sector. Still, the argument is less clear-cut than in the case of county-owned industries. The enterprises owned by supra-county jurisdictions were much less obviously rural than those which fell under the control of China's counties. They were typically located in county towns rather than in ordinary zhen, employed few temporary workers, contributed little (if anything) to the county budget, used more sophisticated technologies, and produced outputs which were usually unrelated to agriculture. For all that, it is hard to regard these enterprises as having much in common with the large and relatively sophisticated enterprises which were the mainstay of industrial production in China's large cities. They were more similar to the xiangzhen qiye and county-owned SOEs, than to the industrial enterprises of Beijing, Shanghai and Daqing.

It therefore seems most appropriate to classify all SOEs and COEs based within county boundaries—irrespective of level of ownership—as rural enterprises. The decisive factor here is their treatment in the Chinese national accounts. For example, the provincial statistical yearbooks published in the 1980s and 1990s provide data on industrial output by county, and these data include the production of all enterprises located within the county boundaries. Similarly, the Population Census data for 1982 on industrial employment within counties cover all types of industries, irrespective of their ownership. It therefore appears reasonable to follow suit, and to put together all the industries located within a county together, rather than to distinguish between them on the grounds of ownership.

A4.3. The Role of the Hukou

The main disadvantage of classifying COEs and SOEs located within a county rural enterprise is that the approach re-classifies urban citizens residing in the county (i.e. those with a non-agricultural hukou), and working in the COEs and SOEs, as rural.10 The same problem applies to those working in zhen enterprises and also having a non-agricultural hukou. This does some violence to Chinese realities because these employees are officially urban (unless they were temporary contract workers). These employees are included in the urban surveys (chengzhen) of income and expenditure because of their hukou status rather than their place of residence.11

The jurisdictions principally affected are county towns where the bulk of the county population with a non-agricultural hukou resided.12 In Wushan county (Sichuan), 56 percent of the non-agricultural population lived in the county town in the mid-1980s (Wushan XZ 1991: 100). In Santai, also in Sichuan, the figure for 1987 was very similar (p.353)

Table A4.3 The non-agricultural population living in China's towns, end 1984


Total town population (millions)

Non-agricultural town population (millions)

Non-agricultural share (percent)





































All China




Source: Gong'an bu (1985: 1).

(Santai XZ 1992: 129). In the more prosperous county of Kunshan in Jiangsu, the non-agricultural population residing in the county town was 62 percent of the county total (Kunshan XZ 1990: 141). Or, to take a Zhejiang example, 21,152 out of Tonglu county's non-agricultural population of 55,490 (38 percent) in 1985 lived in the county town (Tonglu XZ 1991: 81).

Nevertheless, it was rare for more than a minority of the population living in county towns to have possessed a non-agricultural hukou in the mid-1980s.13 The Sichuan basin was an exception; there, the non-agricultural population exceeded 75 percent of the population of the county town in virtually every county. The proportions were also high in peri-urban Shanghai and in Guangdong. In North China, however, the figures were much lower. In none of the county towns of Xingtai prefecture in Hebei (the prefecture was divided into 14 counties at the end of 1984) did the figure come even close to 50 percent. Of Shijiazhuang prefecture's 15 counties, the non-agricultural ratio exceeded 50 percent only in Shulu (which was one reason why the county was upgraded to a city in the 1980s). In Jiangsu's prosperous Wujin county, fully 64 percent of those living in the county town at the time of the 1990 Population Census had an agricultural hukou (Wujin pucha 1992: 12), though this is rather misleading because Changzhou city was in effect the county town for Wujin and therefore home to most of those without an agricultural hukou.

The national picture is apparent in Table A4.3, which lists the non-agricultural population of all towns in a selection of China's provinces at the end of 1984, and which (p.354) shows that the non-agricultural population was in a minority in China's towns in the early 1980s. These data are particularly useful because, although they cover more than just county towns, those jurisdictions designated as urban in 1984 were much more obviously ‘urban’ than the new creations (jianzhi zhen) of the late 1980s. In other words, even the most obviously ‘urban’ settlements within China's counties in the early 1980s were not really urban, at least in terms of the hukou status of the bulk of their population. In the absence, therefore, of large truly urban settlements within Chinese counties, it is reasonable to regard counties as predominantly rural and, by implication, the industrial enterprises located within their boundaries as rural also.

A4.4. Summary

For these various reasons, the approach adopted in this book is to adopt a wider definition of rural industry than is the norm in the literature. Instead of using either the CBE or TVE definition, I classify all Chinese jurisdictions designated counties (xian) as rural, and designate all county-level shi (cities) as urban. It follows that all types of industrial enterprises operating within a Chinese county, irrespective of ownership and size, should be regarded as being rural industries.

Two final observations. Firstly, any distinction between urban and rural areas, whether for the Maoist or post-Mao eras, is ultimately arbitrary. In the Asian context, it has been argued by McGee and others that modernization in many rice-based regions has not led to the ‘classical’ Western process of urbanization. Rather, the distinction between urban and rural areas has been blurred. Small towns have sprung up instead of urban agglomerations, the economically-active members of individual households are engaged in both farm and industrial production, and much industrial production is seasonal. The phenomenon has been labelled as kotadesasi, and it is argued to be widespread in parts of China, and especially in Jiangsu and Guangdong (Ginsburg et al. 1991; Ho 1994); the Chinese phrase litu bu lixiang (‘leave the land but not the countryside’) aptly summarizes the process.14 Precisely because of the existence of kotadesasi, we need to be aware that variations in definition might lead to different results and, as a corollary, to different conclusions on the impact of Maoist programme upon rural industrial development after 1978. Accordingly, the econometric analysis in this book is subjected to sensitivity analysis on this point to see if the inclusion of officially-designated cities in the samples influences the conclusions to be drawn.

Secondly, although any definition of urban and rural cannot be time invariant, it must recognize the on-going process of urbanization which has taken place across China since 1949. For example, many county-level jurisdictions were upgraded during the Maoist era to city status. This upgrading particularly affected areas picked out for Third Front construction (such as Panzhihua in Sichuan or Shiyan in Hubei), or for the development of port or oil-refining facilities (such as Maoming in Guangdong). Should such up-graded jurisdictions be classified as rural (their status at the beginning of, or at least early on in, the Maoist era) or as urban (their status at the close of the 1970s)?

There is no straightforward answer but, given that our interest lies primarily in the extent to which industry developed in rural areas, it makes sense to use the initial status of (p.355) a jurisdiction to define the nature of the industrialization process therein. That of course still leaves open the question of ‘initial status when’? However, given that (by any definition) rural industrialization did not really get underway until the late 1950s, and that jurisdictional boundaries fluctuated wildly during the Leap, it is logical to use the 1964 Population Census delineation of jurisdictional status. In discussing Maoist rural industrialization, I therefore tend to use 1964 status to determine whether a process of industrialization is urban or rural. The industrial development of Daqing is thus an urban process; the city of Daqing was only created in 1979 but the oil industry developed in Anda, a jurisdiction officially classified as urban in 1964. Conversely, the industrialization of Panzhihua classifies as rural industrial development because the complex was created on a site which was officially classified as rural in 1964. The 1964 cut-off date does have the important implication that we should regard much of the Third Front construction of the 1960s as a programme of rural industrialization (because most Front county-level jurisdictions were officially rural in the early 1960s). However, that is hardly inappropriate. After all (see Chapter 2), the explicit aim of the Front programme was to bring about industrial growth in relatively backward, inaccessible and predominantly rural areas. It does, however, mean that the way in which we usually think about the scope of Maoist rural industrialization needs to be revised.

At the same time, it make little sense to view the continuing industrialization of jurisdictions such as Panzhihua in the 1980s as a process of rural industrialization. By then, Panzhihua was a city and its industry was manifestly developing in an urban location. On the other hand, it makes little sense to use the jurisdictional categories of the late 1990s. Many counties were upgraded to city status in the late 1980s and 1990s even though upgrading was hardly warranted by the size of their non-agricultural populations. In some cases, of course, the change in designation was justified; the transformation wrought by the growth of the non-farm rural sector was so profound that it made little sense to regard such places as rural jurisdictions any longer. However, such jurisdictions were manifestly rural at the start of the 1980s and that is how we should classify them: we are interested, after all, in the extent and impact of rural industrialization and therefore we need to look at areas which were initially rural, irrespective of their end-period status. Kunshan in Jiangsu is an example of an essentially rural landscape transformed by industrialization, and it is now (rightly) viewed as a city. Nevertheless, its growth since the early 1980s is properly regarded as the transformation of rural to an urban landscape via industrialization. The obvious year to pick for the identification of status is 1982 because that was the year of a Population Census and therefore the data used to identify cities—occupation of the population, hukou status etc.—was reliable. The national-level econometrics employed in this book therefore uses the term ‘rural’ to refer to jurisdictions officially classified as counties at the time of the 1982 Population Census.


(1) However, the industrial output of brigade and sub-brigade industries was included in agricultural production (GVAO), rather than in industrial production (GVIO), before 1984 (zhen or town industry was typically included in GVIO as zhen were regarded as equivalents of communes). Many of the statistical publications of the late 1980s and 1990s have rectified this, and have produced time series data for GVIO which include the output of the brigade sector going back to 1949. But most pre-1988 publications do not, and the same is true of some of the County Records published in the 1990s.

(2) This was not the case in the 1950s, when COEs were less technologically advanced than SOEs. Many of them were created out of pre-1949 handicraft industries (Donnithorne 1967; Riskin 1987: 98), whereas most SOEs were either foreign companies which had been nationalized, or new industries established with imported Soviet technology. Some of the handicraft cooperatives/collectives in turn became the foundation for commune industries during the Great Leap Forward (Riskin 1971: 264–5).

(3) A typical Chinese county of the mid-1980s was divided into xiang (townships), zhen (towns) and a county seat (chengzhen). Townships and towns were in turn divided into cun (villages) and (confusingly) smaller towns (zhen). Kunshan county in Jiangsu, for example, in 1987 comprised 13 xiang, 6 zhen and the county seat (Kunshan XZ 1990: 90). Zhen were typically so designated because they were seen as more ‘developed’ than xiang, and thus prosperous counties tended to have more zhen than poor counties. For example, prosperous Cangnan county in Zhejiang was divided into 19 zhen and 16 xiang in 2000, whereas (poor) Wencheng county was divided into 25 xiang and only 7 zhen (WZTJNJ 2001:15). Sichuan's prosperous Deyang municipality (made up of 5 county-level administrative units) comprised 128 zhen and 33 xiang, whereas poor Liangshan (16 county equivalents) numbered only 74 zhen against its 542 xiang in 2001 (SCTJNJ 2002: 3). At the opposite end of the scale, Yixing in Sunan had small city status by 1996, and this was reflected in its having 31 zhen and only 13 xiang. In neighbouring Xishan city (formerly Wuxi county), there was not a single xiang (Wuxi tongjiju 1997: 115–17). Nevertheless, standard measures of income, per capita industrial output and the proportion of the workforce classified as non-agricultural often show little real difference between the development levels of zhen and xiang. In Yixing, the xiang with the highest GDP per head was more prosperous than 26 of the 31 zhen (Wuxi tongjiju 1996: 115–16). Evidently, then, the difference in status between the two types of jurisdiction was determined by considerations that were not purely economic.

(4) A siying (private) enterprise is one employing more than seven workers. A geti (individual) enterprise is so classified if it employs seven workers or less. Both types are of course private enterprises; the difference between them is only one of size.

(5) In some counties, these sidelines were of great importance. A case in paper is handicraft paper production in Jiajiang, which provided about half of all rural industrial jobs in the late 1990s (Eyferth 2004: 78).

(6) At the time of the 1982 Census, there were 236 cities and 2,136 counties (Population Census Office 1987: xv). By the end of 2001, there were 2,053 counties and 393 county-level cities in China, reflecting both growing urbanization and the re-designation of some counties as cities (TJNJ 2002: 3).

(7) Many of China's state and collective industries are self-evidently urban enterprises. The largest state-owned industries were located in cities, subordinated to the governance of ministries, provincial and municipal governments, and were a world removed from the rural sector. Much the same applies to enterprises owned by urban collectives (chengzhen jiti suoyouzhi), many of the enterprises so classified being little different in terms of operation, location and governance to their cousins in the state sector. It is the treatment of enterprises located within county boundaries that creates most of the debate (though, as the quotation from Wong makes clear, some would argue that there is a case for also treating prefectural SOEs and COEs as rural enterprises).

(8) In Shulu county studied by Blecher and Shue (1996: 110), there were more temporary contract workers than regular workers in county enterprises in 1976 and even in 1990 temporary workers made up over 25 percent of the regular workforce. There were also a large number of temporary contract workers in southern Jiangsu. In the town of Zhenze in Wujiang county, about 20 percent of the workforce were ‘peasant workers’ (Fei 1986: 200).

(9) In Jiangsu's affluent Kunshan county in 1987, for example, there were 205 companies operating jointly with Shanghai enterprises, 21 cooperating with enterprises based elsewhere in China and five joint ventures with foreign funding (Kunshan XZ 1990: 262–3). In poor Wushan county in Sichuan, five of the 18 SOEs were run by higher-level jurisdictions in 1985 (Wushan XZ 1991: 193). In Santai, also in Sichuan, three of the 52 SOEs were centrally-run; these three employed 3,599 workers (7 percent of all county industrial workers) and produced about 15 percent of gross output value in 1987 (Santai XZ 1992: 454, 460–1).

(10) A further controversial aspect of the equating county with rural and city with urban is that farmers working in (say) Shanghai classify as urban citizens. But in many ways this makes sense; they possessed an urban hukou and their lifestyle was far more urban than that of farmers working in (say) Henan.

(11) These surveys distinguish between households resident in large cities (chengshi) and in county cities (xiancheng), but all are classified as urban to distinguish them from rural (nongcun zhumin) households.

(12) Of course the reclassification of xiang as zhen did not affect the status of citizens living within the affected jurisdictions. Their status attached to their person and depended upon their hukou, rather than their place of residence. For that reason, peasant migrants from zhen working in (say) Beijing and Shanghai continue be classified as peasants unless their hukou status is changed.

(13) For counties taken as a whole, the non-agricultural population was only a small part of the total population in the 1980s and even in the 1990s. In Santai in Sichuan, for example, the non-agricultural population (fei nongye renkou) was 6.1 percent in 1987 (Santai XZ 1992: 125). In Wushan, one of Sichuan's least prosperous counties at the time (it is now part of Chongqing), the corresponding figure was 5.4 percent in 1985 (Wushan XZ 1991: 100). The average for the five county-level jurisdictions within Suzhou in 1995 was 19.4 percent (Suzhou tongjiju 1996: 53). In the four relatively prosperous counties of Changsha municipality in Hunan, the non-agricultural population accounted for 8.2 percent of the total in 1990 (HNTJNJ 1991: 309).The figures were substantially higher by the end of the 1990s. But of course this is the very reason why rural industrialization has attracted so much interest, and in any case the pace of urbanization did not justify the wholesale re-classification of xiang as zhen. For our purposes, the issue is how to classify industries and populations at the beginning of the 1980s. If we wish to argue that rural areas have been transformed by industrialization, we need to be clear that they were rural in the first place.

(14) Kota (city), desa (village), si (process) is Indonesian in origin (Ginsburg et al. 1991: xvii). Kotadesasi and desakota are used inter-changeably in the literature.