Importing export zones: processes and impacts of replicating a Chinese model of urbanization in rural south India

ABSTRACT This article examines the development of a south Indian “industrial city”, modelled on a Chinese counterpart. The privately-operated city is a key example of India’s new-style Special Economic Zones (SEZs). These represent a national shift, motivated by China’s SEZ success, from enclave-style “export processing zones” to integrated townships. Drawing on fieldwork from 2018, we compare three issues with the Chinese “model”: the city’s establishment and spatial planning; the gendered hiring practices of its firms; and its incorporation of local villages. These three aspects represent different mechanisms through which the “model” is translated into the Indian context, by different actors, with different outcomes. In line with policy mobilities scholarship, we argue that models cannot be straightforwardly replicated, and we call for more attention to the outcomes of attempted replication, in particular the urban development implications of selective, complex, multi-level adaptation of a Chinese “model” and its interaction with local Indian contexts.


Introduction
From the 1960s, the developing world began a shift toward export-oriented industrialization (Farole & Akinci, 2011). Countries with abundant labor and limited foreign exchange created export processing zones, which played a significant role in attracting foreign direct investment (FDI) and enabling a system of global production to become paramount. Offering, typically, regulatory regimes with exceptions to national tax, labor and customs laws, confined geographic areas with enhanced infrastructure and transport links, and governance dedicated to enabling business, they became privileged economic spaces for manufacturing (Farole & Akinci, 2011;Neveling, 2015). Among many such zones worldwide, China's "Special Economic Zones" (SEZs) from the 1980s on stood out for their stunning economic success and their transformative impacts on both economy and society.
China's SEZs facilitated the country's transition from an agricultural to an industrial economy, by inducing the migration of millions of former farmers, who, working under new and often exploitative labor regimes, transferred economic gains from coastal areas to the rural interior. The success of this model attracted many imitators across the developing world from the 1990s on. More recently, China itself has become increasingly comfortable suggesting its approach provides a useful example (Jiang, 2018), while the country's growing international engagement provides a vehicle for spreading its practices (Knoerich & Urdinez, 2019). Sharing similarities with China in terms of population size, economic growth rate and agricultural sector size, India offers an interesting case study of how China's "model" is reproduced elsewhere. Since 2005, it has attempted to establish new Chinese-style SEZs and reinvigorate older export zones based on China's model. However, India's different political, social, cultural and temporal contexts raise questions about the feasibility of adopting China's SEZ experience, itself based on a set of global, national and local factors which produced distinctive developmental outcomes.
This article examines this problem of attempted replication of China's SEZ experience in India, by focusing on the development and impacts of one anonymized Indian SEZ. This zone, designated a new "industrial city", was directly modeled on a Chinese counterpart and hosts Chinese firms. 1 Drawing on fieldwork conducted in the zone in 2018, the article addresses the question of how a Chinese SEZ "model" interacts with specific Indian local contexts to produce a set of outcomes both similar and dissimilar to the Chinese "original". Through a grounded empirical examination of the SEZ model, it develops three key lines of comparison, representing different mechanisms of policy mobility: planning (the zone's establishment and spatial configuration); recruitment (firms' hiring practices within the zone); and urban integration (incorporation of villages). We examine how these three different elements of the "model" are mobilized by different actors; observe their mutation in response to local political, economic, social and cultural contexts; and analyze the new forms of urban experience this produces. Through in-depth examination of this case, we demonstrate how aspects of a Chinese "model" are reproduced -both intentionally and unintentionally -in the Indian setting. We conclude that models are not static templates to be copied, but historically contingent, context-specific, dynamic assemblages.
Through this line of examination, we address a relatively neglected issue in the literature on Indian SEZs. This has focused on their economic potential (Tantri, 2016), the role of local states (Cross, 2014;Kennedy, 2014;Levien, 2013), and their effects for dispossessed locals (Sampat, 2010), but there is little which specifically compares the Indian experience with the Chinese "model". That which attempts comparison highlights how Indian SEZs have failed to "replicate" China's economic success, but does not investigate the complex variety of ways the "model" is emulated or how it interacts with important local contexts (Aggarwal, 2010;Palit & Bhattacharjee, 2008). Nor do such comparisons examine whether Indian SEZs have reproduced unintentional side-effects of the Chinese developmental path. Understanding these mechanisms and their outcomes is of global significance given the enthusiasm of developing-country policymakers for the Chinese SEZ "model".
The article first provides a critical examination of the literature on policy mobilities and urban "models". After summarizing the methods, it reviews China's SEZ "model", before examining India's attempts to emulate it at the national level. We then analyze the three lines of comparison: planning, recruitment, and urban integration. The conclusion returns to policy mobilities to show how we might better approach "models", with their complex array of actors, components and outcomes.

Policy mobilities and urban "models"
The field of policy mobilities, developed by political and urban geographers over the last decade, positions itself as an alternative to political science's policy transfer approach (Benson & Jordan, 2011). Whereas policy transfer conceptualizes the movement of policies in terms of the careful selection and application of best practices by rational decision-makers, policy mobilities research provides a more geographically-sensitive approach with emphasis on socio-spatial contexts and the relationality of policymaking (McCann & Ward, 2012;Peck & Theodore, 2010;Temenos & McCann, 2013). Such work is often based on detailed reconstructions of the movement of specific policies, ranging from business improvement districts to sustainable urbanism or conditional cash transfers. It highlights that policy adoption is a grounded process, that its impacts depend on context, and that policies change and mutate as they travel (Peck & Theodore, 2012;Ward, 2006).
One strand of the policy mobilities literature is concerned with urban models, and the actors who facilitate their movement. Policies are generally not imposed from outside; instead, cities go through a process of "selective policy emulation" in which they adopt certain principles from elsewhere through a variety of channels (J. Zhang, 2012). Groups or "assemblages" of principles have become associated with paradigmatic city models, increasingly used to inform urban development policy (McFarlane, 2011). Such models were originally based overwhelmingly on European and North American cities, but more recently Asian cities have provided alternatives (Robinson, 2002;Roy, 2016;Shin, 2019). The export of Asian models is facilitated through a network of state actors as well as crossborder networks and consultants, who contribute both expertise and funding (Shin, 2019;Shin et al., 2020). Studies of Korean-funded satellite cities in Phnom Penh, Kuala-Lumpurstyle high-tech parks in Hyderabad, and Singaporean urban greening in Dalian, among many others, suggest that this "intra-Asian urbanism" is a growing phenomenon (Bunnell & Das, 2010;Chua, 2011;Hoffman, 2011;Percival & Waley, 2012).
However, when a model is adopted elsewhere, it is adapted. The literature reveals several reasons for this: actors adopting the model prioritize aspects they find attractive (McCann & Ward, 2012); the model must be transplanted into different conditions (Ong, A. 2011); the model is more an abstract ideal than a realizable plan (Shin, 2019); and there may be different actors at different levels involved in translating the model (Bok & Coe, 2017). This last point is particularly important: while much literature has examined policy mobilization through the efforts of a small group of elite, typically state, actors (Peck & Theodore, 2012), others, including non-elite state and non-state actors, acting through informal or ad hoc networks, may also play a crucial role (Baker et al., 2020;Cohen, 2015).
Their role may help to explain the lack of connection between the existing literature on Chinese-style SEZs, including transnational SEZs, and that on policy mobilities. Although SEZs have been examined extensively both as an economic strategy for boosting exports (Farole & Akinci, 2011;World Bank, 2015) and as a governance tool employing "zoning technologies" to create distinct regimes within a national landscape (Ong, 2004;Sood & Kennedy, 2020), scant attention has been paid to how, why or with what results components of an SEZ model move from one setting to another. Translating industrialization into new contexts, and integrating this with urban development, is often driven not at the level of national government policy, the focus of most policy mobilities scholarship, but by private developers, firms and their employees (Bok & Coe, 2017). These actors may have no intention of engaging in policy mobility, but nonetheless in pursuit of profit translate components of a model into new contexts. We extend this line of analysis to suggest that some actors adapt the model intentionally, some unintentionally; and sometimes unwanted or unobserved aspects of the model may be reproduced and transfigured in the new context. This complex interplay of factors produces messy and potentially quite different outcomes for urban development compared with the "original" (itself historically contingent, and producing unintended effects). This study draws attention to the multiple ways in which the Chinese SEZ model is translated into and interacts with specific Indian contexts, to produce developmental outcomes both reminiscent of the Chinese urban experience and yet strikingly different.

Data and methods
Fieldwork for this project was conducted in December 2018 in an anonymized privatelyowned "city" in southern India. The city was founded in the 2000s and contains a large SEZ as well as a domestic industrial zone where global firms manufacture for the Indian market. The city hosts over 100 companies from around 25 countries. It is unusual among Indian SEZs for its size and rural location, as set out in Table 1 below. It was selected as a case study site not for any claim to representativeness, but for two key reasons: first, because it stands as an example of the new Indian approach to establishing Chinese-style zones, encompassing residential, healthcare and educational provision, retail and leisure activities as well as factories; second, because it hosts a significant number of Chinese investors as well as those from elsewhere. This facilitated examination of how China's "model" is translated into the Indian context, including the role of firms with experience of manufacturing in China.
In total, 74 semi-structured interviews were conducted with zone managers (5), firm managers (5), white-collar employees (4), local government officials (4), labor contractors (3), local villagers (20), migrant hostel operators (4), and both local (15) and migrant (14) assembly-line workers. Zone and firm managers and local labor contractors were Source: Authors, based on (Farole & Akinci, 2011;Jenkins et al., 2015;Wong, 1987). approached in their offices; villagers were identified through visits to villages now within the city boundaries; and migrant workers and hostel operators were approached at local hostels. Workers were those currently or formerly employed in Chinese or other firms, and were mostly introduced by fellow villagers or hostel-mates. Villagers not employed in factories were approached in their villages and included based on their availability and willingness to participate. Interviews with zone and factory managers, officials and whitecollar workers were in English, while those with local and migrant workers were conducted with the help of research assistants fluent in local languages. Data drawn from these interviews as well as from publicly available materials about the city, documents provided by zone managers, and observations conducted in the local area form the basis for a grounded empirical examination of the local Indian interpretation of China's SEZ model. The three lines of comparison emerged inductively through a process of category development, involving repeated examination of the data for phenomena reminiscent of Chinese SEZs, from explicit mention of Chinese inspiration to actual, intended or implicit similarities with Chinese practices. The resulting lines of comparison encompass three different mechanisms of the "model's" mobility through different actors and processes, but do not represent all possible avenues through which a Chinese "model" may be adopted in India.
The name and location of the city and its incorporated villages are not specified, and all firms and individuals are anonymized to protect the fieldwork site and its inhabitants, most of whom agreed to speak on condition that they not be identified. Although this brings drawbacks in terms of presentation of contextual information -for example, certain visual data, zone materials, policy documents and company information must be summarized rather than presented in the original -it ensures as far as possible that neither individuals nor communities are exposed to potential negative consequences, such as retaliation in workplaces, accommodation or villages; damage to business operations or reputation; or personal exposure. Although anonymization limits direct presentation of data, it does not limit the depth of analysis, which is based on fieldwork results, primary documents and a rich set of secondary material.

The "model": SEZs in China
The core features of China's original SEZs have been enumerated, usually including the following: a large, geographically-delimited, physically-secured area of former rural land; governance by comprehensive national legislation, with local-level autonomy to develop laws and administer zones; benefits for foreign investors, including financial incentives, exemptions and more relaxed labor regulation; and labor-intensive manufacturing, employing primarily young, female rural migrants (Farole & Akinci, 2011;World Bank, 2015). Scholars have compared Indian SEZs with this Chinese "model", pointing out where India has fallen short, and examining the economic implications (Aggarwal, 2010;Palit & Bhattacharjee, 2008). In this scholarship, models are viewed as rigid prototypes fixed by the originator -here, China. This perspective ignores how models are translated into different contexts by various actors over time, a process in which the perspectives of "translators" and others in the host country are key. Accordingly, this section focuses less on listing key features of Chinese SEZs, and more on understanding how, why and from where they were adopted and adapted by China in the 1980s.
SEZs were established in China to overcome economic and technological weaknesses following 30 years of relative isolation under Mao Zedong's leadership. Taiwan and South Korea, both successful in gaining capital, technology and economic growth through export processing zones in the 1960s-70s, presented an attractive "model" for countries with abundant labor and limited foreign exchange (Neveling, 2015). Complicating the model's origins further, Taiwan and South Korea had themselves looked toward the colonial entrepots of Singapore and Hong Kong for inspiration (Sit, 1988), while the origins of contemporary zones, as "spaces of exception" to tax, labor and customs laws (Ong, 2006), are usually traced back to Shannon, Ireland, in 1958. These examples raised ideological dilemmas for communist China, yet with the new term "special economic zone" leaders argued that these would not be capitalist export processing zones, but model cities with diverse industries including tourism and real estate (Wong, 1987). In 1980, Jiang Zemin, then an official and later President, was dispatched to Shannon to learn how the Irish zone worked, and China's first four SEZs -Shenzhen, Zhuhai, Shantou and Xiamen -were established along China's coastal periphery, not only to attract investment from Hong Kong and Taiwan, but also to ensure they could be easily erased if they failed Neveling, 2015).
From the start, Chinese SEZs were designed as experimental zones, sites of transformation as well as production. Moving beyond their economic impacts to their sociopolitical, Ong (2006) famously conceptualized China's SEZs as a "neoliberal exception" to China's planned economy, segregated from the rest of the nation by "zoning technologies". These created distinct governance regimes within the national polity, and rescaled state power through "graduated sovereignty" which accords different rights and privileges to different population segments depending on their relation to the market (Ong, 2004). Use of similar 'zoning technologies' has since been explored in the context of export zones in other parts of Asia, including India (Sood & Kennedy, 2020;Tsai, 2015). Yet the "exceptionality" of China's SEZs has been challenged. Cartier (2018), for example, points out their limited autonomy under the party-state; the decline of Chinese central planning more widely after the 1970s; the array of other "zones" (qu) established during the 1980s-90s; and the existence of differential scales of regulation more broadly in China. Other scholars of SEZs globally have criticized "neoliberal exceptionality" for overlooking the permeability of neoliberalisation, the ubiquity of migrant precarity, and the level of internal stratification within SEZs (Cross, 2010;He & Chang, 2020;Park, 2017).
While acknowledging the distinct governance mode of China's SEZs and their regulatory difference from other parts of China in the early 1980s, we also wish to highlight their rootedness in existing Chinese party-state relations, their debt to earlier socialist urban "models", and their critical role as policy test-spaces (Bach, 2017). Understanding SEZs in this way allows a focus on the "model" as a contingent assemblage emerging over time. We thus characterize China's SEZs here primarily as "experimental" rather than "exceptional". As a temporal threshold where reforms could be tested before being extended nationwide, as well as a spatial threshold mediating between China's economic space and that of other countries, China's SEZs became examples for the rest of the country to follow as well as for the rest of the world to observe China's development.
Their flagship was Shenzhen -a new centrally planned industrial city with a full range of urban functions, as well as an export processing hub, populated by "civilized" citizens (Bach, 2010). By 1985 the Chinese government had spent six billion yuan (USD 2.1 billion) integrating over 20 residential, commercial, industrial and tourist areas in the new city (Bach, 2017). The preferential treatment for foreign investors, new possibilities for joint ventures, and the experimental contract labor system, which led to massive in-migration, worked to ensure rapid investment, in particular from neighboring Hong Kong. From 1986, capital was channeled specifically to the development of export-oriented industry, particularly to attract large, high-tech foreign enterprises (Wong, 1987). Many multinational corporations set up manufacturing bases in Shenzhen, employing rural women recruited first from Guangdong's former production brigades, and then from interior provinces (Andors, 1988). By 1989 there were more than one million temporary workers in the zone, 80% of whom were women (Sklair, 1991).
The experiment was a great economic success: Shenzhen's trade balance reached USD 1.64 billion by 1994 (Shenzhen Statistics Bureau [SSB], 2008). It is now a megacity of around 20 million people and 500 km 2 , one of China's principal import-export hubs and a globally-leading manufacturing center. As a result of its success, in addition to the initial four SEZs, China now has a wide range of other economic "zones" of different shapes, sizes, locations and nomenclatures, including Economic and Technological Development Zones, Free Trade Zones and others (Aggarwal, 2010;Ngo et al., 2017). They are estimated to have contributed 22% of China's GDP, 45% of FDI, and 60% of exports, accelerating nationwide industrialization, agricultural modernization, and urbanization (World Bank, 2015). These newer zones are mostly attached to existing cities, without physical boundaries, and with a range of governance structures, suggesting substantial evolution within China from the SEZ "model" of the 1980s.

Emulating the Chinese SEZ "model" in India
China's adoption of the zone model addressed a need to acquire foreign capital and technology after years of Maoist isolation. These specific circumstances and the experimental nature of Chinese SEZs notwithstanding, their economic success meant that other developing countries looked to emulate them to boost their own economies. In this section, we explore how and why Indian policymakers selectively adopted (what they believed to be) key features of the Chinese model. Our aim here is not to compare and contrast Indian and Chinese SEZs along a fixed set of criteria (for this, see Aggarwal, 2010;Kundra, 2000); rather it is to understand how and why Indian actors borrowed aspects of China's experiment, and the context into which they were translated, often producing different outcomes.
The Indian state aspired to imitate China in expanding manufactured exports and employment by setting up new SEZs and converting older export processing zones (EPZs). India's EPZs predate China's SEZs, with Gujarat establishing the first in 1965, followed by Bombay in 1973 -the first such zones in Asia (Kennedy, 2014). EPZs did not enjoy fiscal or customs incentives, however; FDI regulations were strict; and bureaucratic procedures complex (Tantri, 2016). In 1994, the Indian Council for Research on International Economic Relations sent a "fact-finding mission" to China to identify "useful features of Chinese zones that could be adapted to Indian conditions" (Cross, 2014, p. 37). The first official mention of SEZs in India followed Commerce Minister Murasoli Maran's visit to Shenzhen in 2000, which he described as "an eye-opener" (Palit & Bhattacharjee, 2008, p. 88). Borrowing the Chinese term "SEZ", Maran initiated new rules for the establishment of private zones (Majumder, 2003). The 2000 Export-Import Policy converted existing EPZs, typically modest industrial parks, into SEZs, envisaged as encompassing the full array of facilities that make up a city -housing, hospitals, schools and retail developments -occupying up to half the total area (Government of India, Ministry of Commerce and Industry, 2005). These SEZs would operate through "selfcertification" on tax-exempt transactions, and be governed by new comprehensive legislation (Singh, 2009).
The 2005 SEZ Act was described explicitly by a government report as "a growth catalyst, on the model of the SEZs in China" aiming to "help India replicate the Chinese success story of rapid industrialization" (Parliament of India, 2007). It encompassed many policy domains: not only trade and investment, but also radical deregulation, infrastructure creation and tax regime changes. Its official objectives were to promote economic activity, exports and investment, create employment and develop infrastructure (Government of India, 2005). However, it also aimed at overcoming "deficiencies" of India's existing policies, including barriers raised by monetary, trade, fiscal, tariff and labor regulations, and -in contrast to China -at encouraging private investors to take responsibility for developing infrastructure (Aggarwal, 2010;Sampat, 2010). In setting up "self-contained zones" including residential and commercial areas, the developer would be responsible for providing civic amenities, roads, sewerage, housing, utilities, green spaces and education -in essence, taking over the role of the local municipal government (Menon & Mitra, 2009).
The idea of corporate-owned or -operated urban areas in India was not new. Company towns were a feature of Nehru's post-independence development project. Large tracts of rural land were acquired for factories and "industrial townships", with public sector steel towns prominent from the 1950s (Levien, 2013). In both traditional company towns and new SEZs, employers and operators acquired immense power, not only within the confines of the workplace but also beyond. Some of these powers were devolved from state-level governments, for example, authority in urban governance, service provision and even tax collection (Sood, 2015;Vaidya & Dar, 2008); while others represent the gradual extension of corporate control into workers' private spheres such as housing, leisure and daily social reproduction (Goodburn & Mishra, n.d.). The difference between earlier forms of corporate urban development and contemporary Indian SEZs is, according to Levien (2013), that earlier state-led projects aimed at production and enjoyed public legitimacy, whereas SEZ land is expropriated for private capital with the state as broker, amid public opposition. Cross (2014, p. 34) adds that, whereas earlier industrial towns had enabled controlled experimentation with private enterprise in a way which protected state-owned industries from the market, new SEZs were designed to protect private capital from the state.
The role of the private sector is the major feature which distinguishes the Indian approach from the Chinese "model". In China the state was overwhelmingly responsible for SEZs. Village governments transferred rural land-use contracts from village collectives to the state, which redesignated the land "urban" so that the rights to develop it could be sold; nonetheless, the state retained ultimate ownership and responsibility for planning, infrastructure and administration (L. Zhang, 2011). In India, though some investors have partnered with government bodies, the largest share of capital has come from private entities (Kennedy, 2014). The state's role has usually been limited to that of broker -acquiring the land for "public purpose" under India's Land Acquisition laws, often via a parastatal industrial development corporation (Balakrishnan, 2013(Balakrishnan, , 2019Sood & Kennedy, 2020), then transferring ownership to developers. Private developers then typically lease the land to firms on multi-year contracts for construction of their plants. Designated "industrial townships", many SEZs are exempt from the constitutional requirement for elected municipal self-government, and zone developers remain responsible for administration and security (Sivaramakrishnan, 2009). Such use of "zoning technologies" enables parastatal industrial area local authorities to facilitate new bounded spaces essentially governed by private entities. The resulting "regimes of exceptional governance" may thus represent both neo-liberalization and new illiberal forms of territorial control (Sood & Kennedy, 2020, p. 25).
A second major difference in translation of China's SEZ "model" -probably a result of the first -is the lack of emphasis in India on SEZs as a mechanism for promoting strategically located under-industrialized areas. Only Xiamen of the original Chinese SEZs was already urban, whereas India's SEZs are overwhelmingly in developed areas, leading some to interpret SEZ establishment as a form of acquiring valuable real estate (Jenkins et al., 2015). 2 Relatedly, India's SEZs are mostly much smaller than those in China, as Table 1 above sets out. India's 2005 SEZ Act specified a minimum size for sector-specific SEZs of just 100 hectares (10 hectares for those in IT). Only the minority of non-sector-specific SEZs had a size requirement of 1000 hectares (Government of India, 2005). The case study analyzed here is the latter type of zone: large, and located on former rural land in a district with a below-average level of urbanization both for the state and for India as a whole (Census of India, 2001). These SEZs present different features from the mass of small Indian zones, and raise issues not of urban real estate but of dispossessing farmers and incorporating rural settlements, in the manner of China's original SEZs. As in China, this has led to opposition where lands have been expropriated, as well as concern over urbanization of the countryside, and -unlike in Chinafear of privately-developed gated townships inaccessible to locals (Bedi, 2013;Jenkins et al., 2014;Sampat, 2010).
Despite divergent contexts, broad resemblances in policy aims can be observed. In both countries, SEZs are not only a vehicle for economic growth and employment, but have been a politically expedient way of introducing controversial reforms. In India, to the extent that SEZs create new industrial townships, SEZ policy has been a convenient means of overcoming bureaucratic, legal and political obstacles to urban redevelopment, including, for example, the Urban Land Ceiling Act and the Rent Control Act. In fact, such strategic use of SEZs may have been an appealing part of the Chinese "model" to Indian policymakers: in 2005, Nath commented approvingly that the new SEZs would allow massive "rurbanisation" free of the "shackles of the government inspector" (Kothari et al., 2010). In this sense, we may see SEZs in both countries as "experiments" in "exception" to the prevailing regulatory regime.
In both countries, incentives and exemptions are offered to international investors, with labor regulations particularly subject to official or unofficial deviation. Chinese SEZs were used as experimental areas for a new contract labor system (Wong, 1987), and many Indian zones designated "public utility services" to avoid certain provisions of Indian labor law and the scrutiny of the local labor commissioner (Singh, 2009). Whether this is genuinely "exceptional", given lax enforcement of labor laws nationwide, has rightly been questioned (Cross, 2010). Nonetheless, the presentation of many Indian SEZs to investors as offering a uniquely "flexible" labor regime indicates their emphasis on minimizing production costs in order to boost international competitiveness (Singh, 2009). Multinational corporations have thus moved in, transferring practices from their Chinese experiences, including, in particular, the temporary engagement of young women rural migrants in labor-intensive manufacturing (Palit & Bhattacharjee, 2008, p. 18).
Having examined how and why Indian policymakers translated selected aspects of China's 1980s SEZ "model" into the Indian context, with varying results, the following sections turn from national-level policymakers to a range of actors in our case study zone. They analyze three lines of comparison, which aim not to capture the whole of the Chinese experience, but to represent three different mechanisms through which the "model" is translated, leading to different types of outcome: first, the city's establishment and spatial planning; second, the gendered hiring practices of firms based in the city; third, incorporation of local villages into the city. Each line of comparison sheds light on the complex ways in which a "model" is reproduced; suggests a hitherto underexplored mechanism of reproduction; and examines the variety of outcomes, both intentional and unintentional, which arise through the interaction of the Chinese "model" and the Indian contexts.

Planning the city: establishment and spatial configuration
There is evidence, in the way that the Indian "industrial city" was established and spatially conceived, of an attempt to emulate a specific Chinese model. The city was cofounded in the 2000s by a former politician and a business leader originally from the local area. The location was chosen not only for its connection to the founder but also its proximity to ports and other infrastructures. The land was acquired in phases by the parastatal industrial development corporation, invoking the Land Acquisition Act, and handed to the founders over five years. Unusually for an Indian SEZ, and similar to those of 1980s' China, it was established rurally rather than on the outskirts of an existing city. Part of the stated aim was the transformation of the local economic landscape, from a poor agricultural region to an international manufacturing hub no longer reliant on neighboring cities to provide employment. For inspiration, the founders looked to China, visiting several free trade areas before identifying a major industrial zone (hereafter CZ1) as a suitable model of spatial transformation. 3 CZ1 had been established as a flagship joint-venture project between China's central government and a third-country government, spread over a vast area of land in anticipation of industrial and residential expansion. This forward planning, and the nature of CZ1 as a mixed industrial, residential and commercial area, inspired the Indian zone's founders, who hoped that with a similar urban plan they would attract a similar range of high-quality international investors.
To ensure a similar spatial layout, they hired the third-country consulting firm that had designed CZ1 to conduct the master planning of the city: a deliberate attempt at replicating both urban planning and architecture. CZ1 had placed strong emphasis on "scientific planning", providing extensive infrastructure before construction of factories, and strictly dividing the zone by sector and function. Owing to CZ1's location on the outskirts of an existing city, transport, infrastructure and services were well-connected, and CZ1 was part of the host province's urban development agenda, integrated in planning from early on. By the early 2000s, CZ1 itself had well-reputed urban and social amenities, including clearly demarcated commercial centers, education areas, leisure and recreation districts, as well as industrial and high-end residential areas. Local people were resettled into surrounding peri-urban areas, compensated with cash and rooms to let out, and provided with schools, clinics and training centers to assist farmers into urban manual work.
Examining the spatiality of the Chinese and Indian sites serves to demonstrate the limited extent to which an urban model can be transplanted into a different national context. Given China's vast speed of growth and industrialization, spaces are laid out in anticipation of rapid expansion. While India has a similar context in its large population and economic growth, its industrialization and export-orientation do not match China's; in particular, there have been lower levels of FDI in labor-intensive manufacturing, with most investment in tertiary sector activities (Kennedy, 2014). In 2018, the Indian city was therefore spread over a vast terrain of mostly still-unoccupied land, its multi-lane roads almost empty. Leisure and commercial areas were unfinished, a source of complaint for the Chinese firm managers interviewed, who compared the lack of facilities negatively with the early establishment of amenities in China's zones. They highlighted the role of the state in Chinese infrastructural development -in the words of one factory manager, "In China all the basic facilities, like shopping malls and mobile signal, would be here first; it would be mandatory. The government would send a command to establish them. But here, it's democratic . . . the government can't just tell Airtel to set up a tower". Meanwhile, as discussed below, villagers complained that the land they had sold to developers lay still-undeveloped, yet they were deprived of access through it by the internal customs boundary dividing the city's domestic production zone from the SEZproper, which (like Chinese SEZs of the 1980s, but unlike CZ1) could not be crossed without an official pass.
Differences in speed of expansion and provision of amenities were ascribed by Chinese managers to the difference between democratic and authoritarian governments (on the role of authoritarian states in enabling "Asia speed" development, see Shin et al., 2020). However, it was clear that different ownership structures also played a crucial role: the private nature of the Indian zone meant that there was no state-directed infrastructure development. Although the 2005 SEZ Act dictates that the state provide infrastructural services, its responsibility ends at the gates of the zone (Government of India, 2005). The city is thus largely autonomous, lacking local municipal government, and the private developer is responsible for all development and administration of facilities. This may be advantageous for state governments, gaining infrastructural and urban development despite inadequacy of public monies (Sood, 2015(Sood, , p. 1362. However, the outcome is that, with no state investment inside the SEZ -and the generally more limited state ability to mobilize large development projects in India than in China -the pace of SEZ expansion is considerably slower. Concerned about profits and losses, the Indian zone developers had to expand in line with demand from investors, rather than broader policy ambitions for local development. In the eyes of the SEZ managers, slower progress thus indicated prudent financial management.
Another major spatial difference between the sites is the remoteness of the Indian location, unusual among Indian zones. Although SEZ development in 1980s China focused on strategically-located underdeveloped areas in accordance with central state priorities, the proliferation of industrial zones since has included many adjacent to existing urban areas, established by provincial and city governments, including CZ1. By contrast, the Indian case study site is located on former agricultural land, 80 km from the nearest large city and 30 km from the nearest town, with connecting roads in poor condition -another source of complaint for Chinese firm managers. This location was vaunted by zone managers in terms of its access to ports and airports, but makes urban integration far more difficult than in the CZ1 case. Nor is it akin to China's original SEZs, developing rapidly in strategic peripheral locations through huge state investment. Despite superficial similarities in internal spatial configuration, then, differences in ownership, location and infrastructure investment mean that the Indian industrial city is fundamentally different both from the immediate model of CZ1 as an industrial zone connected to a major city, and from the original Chinese SEZs like Shenzhen.

Recruitment in the city: firms and local and migrant labor
While the city's planning represents deliberate emulation of a specific Chinese model by the zone's founders, the recruitment practices of its multinational firms enable a different kind of translation of China's experience. China's SEZs have long been staffed by "working little sisters" (dagongmei), migrating from rural areas to work in low-skilled, labor-intensive industries including garments, toys and electronics (Pun, 2005). The reproduction of this gendered pattern in the Indian industrial city provides an example of how corporate actors contribute, in less explicit fashion, to the reproduction of a development model, again with different effects from the "original". Of the city's 41 multinational firms, five were Chinese-owned, with the rest from elsewhere in Asia (15), Europe (13), the US (7) and Africa (1). Many had extensive previous manufacturing experience in China, and some play major roles in Chinese supply chains. Although no manager interviewed explained this as a deliberate attempt at replication, both Chinese and non-Chinese firms recruited young Indian women to factory work in a manner reminiscent of recruitment in China's SEZs since the early-1980s.
The literature on young migrant women's labor in Chinese SEZs has examined the ways in which women are hired and managed by firms, and the formation of new social identities drawing on both rural-urban and gender inequalities (Lee, 1998;Pun, 2005). While young women's factory labor is by no means original to China, the gendered regime of precarious employment in export-oriented multi-national manufacturing has been elevated to the status of a "model" in the Chinese case (Pun & Smith, 2006). Gendered hiring practices used in China by both Chinese and non-Chinese firms were replicated in the Indian city, but, in conjunction with India's differing labor regulations and socio-cultural setting, neither the impacts on labor management nor on the women themselves were straightforwardly reproduced. As the Chinese SEZ labor model interacts with the Indian context, there emerge new forms of gendered Indian urban experience. This suggests a need for greater attention to the implications for gender -and indeed other societal arrangements -when examining policy mobilization from one sociocultural context to another.
Women aged 18-23 were preferred as Chinese SEZ factory workers for several reasons. First, they were seen as more docile than men, and unlikely to object to low wages (Pun, 2005). In the Indian case study site, too, this claim was articulated by firm and zone managers, and by workers themselves. Young men complained that it was difficult for them to find work, since they were viewed as potential troublemakers by factory managers, unlike young women from their villages. As one man put it, "they only want the girls". Another recounted how a man from his village had been rejected for skilled factory work: "because he was more educated, they thought he would form groups and create problems. They have this suspicion of local men". The recruitment of an overwhelmingly female labor force in many labor-intensive factories can thus be seen in both countries as a strategy to reduce the likelihood of unrest, considered important for attracting FDI. Although India's national labor laws in theory apply equally in SEZs, in practice many tactics are used to prevent industrial disputes, including the designation of SEZs as "public utility services" rendering strike action near impossible, as well as the delegation of the usual role of labor commissioner, responsible for dealing with violations of Indian labor laws, to the SEZ's own development commissioner (Singh, 2009). In the case study city, zone managers stressed the lack of labor unions and the benefits for investing firms, saying "We have no troubles like that here, our workers are happy. In India, a union means strike; when firms see no union, it means no strike".
Second, the nature of the labor meant it was seen as particularly suitable for women, both by zone and firm managers and by the local community. In the Chinese case, gender stereotypes about women's "nimble fingers" have long been used to justify hiring women on assembly lines, as well as the idea that women more willingly accept tedious, repetitive work (Lee, 1998). This again was frequently mentioned by managers in the Indian city: rural women provided "nimble fingers", "good attention to detail", and were "harmonious workers" even when the work was "very monotonous". An additional factor in the Indian case was that some work was in traditionally female roles such as sewing: this was highlighted as uncontentious work, to which husbands and fathers would not object. The much lower rates of female workforce participation in India, compared to China both now and in the 1980s (Klasen & Pieters, 2015), make it more likely that male family members would reject women's engagement in paid labor without extensive assurances about the nature of the work as well as working environment and accommodation.
The Indian zone managers, as well as some firm managers (both Chinese and non-Chinese), were proud of the social changes they believed would arise from women's employment. The president of the firm operating the zone highlighted the advantages as he saw it: women moved from home-based labor into factories; older, married village women could find sewing positions and contribute to household income; younger girls became earning members of their families, reluctant to leave jobs even on marriage, which was often later than if they had not worked; overall women gained more say in household spending, and would invest in the education of daughters to give the next generation even greater employment opportunities. These positive outcomes are also referenced heavily in the city's promotional literature. However, research into the impacts of women's labor market entry provides a mixed picture of scarce employment alternatives, unequal social status, and undiminished domestic burdens -including even in China, where there may be reasons to believe young migrant women make greater gains in autonomy compared with India (Goodburn, 2020).
Although many local women are employed in the zone, the demand for factory laborers has necessitated encouraging thousands of women aged 18-23 to migrate from across the state. There is little inter-state migration (a major difference from the Chinese "model") because an agreement with the state government prioritizes withinstate recruitment. In the Chinese context, in the early 2000s young women would typically "go out" for three or four years before returning to villages to marry, and their migration experiences allowed them greater say in spouse selection and, perhaps, other family decisions (Fan, 2007). Later research suggests a growing desire among young women to remain in the city, marrying at least another migrant if not an urban husband (Fang, 2012;Goodburn, 2015). In the Indian case, however, longer-term settlement in or near the industrial city appears extremely unlikely, and any gains to young women's autonomy seem largely offset by the repressive workplace and accommodation regime to which they are subject.
Whereas Chinese migrant workers were typically accommodated in factory dormitories, enabling a 24-hour supply of labor (Pun, 2005), in the Indian city migrant women were housed in privately-run hostels outside the city gates. These hostels were subcontracted by firms employing the women; places were allocated to workers in rooms sleeping 5-10 women on floor mats, with fees (typically Rs 800 a month) deducted from wages at source. Buses collected the women from hostels before each shift, and returned them when the shift was completed. All food was provided in the hostel, and women were not allowed to leave without a permit and permission from her employer. The exception to this rule was a weekly excursion in small groups, with a warden, to buy essential items such as toiletries. Other reasons for leaving the hostel, even to attend local temples, were refused.
The reasons given by hostel managers for these restrictions on women's freedom of movement centered around the idea of women's "safety" (physical, sexual and moral). Women were "young", "not from this place" and "at risk of getting up to who-knowswhat". One hostel manager confirmed restrictions were required by firms, and that he monitored women's movements to report to employers. Families would not allow their daughters to migrate for work if their safety could not be guaranteed in this way, and thus controlling their movement allowed firms to ensure adequate supply of young, female labor. One young woman migrant pointed out another way in which curtailing movement ensured labor supply: "it is also so that we do not leave for our villages after taking our wage". These repressive conditions, in which women are shuttled only between hostel and factory, mean that Indian migrant workers do not experience the kinds of emancipatory effects Chinese migrant women may experience. This seems paradoxical, since Chinese migrants live on-site and are subject to excessive overtime through the "dormitory labor regime" (Smith & Pun, 2006), yet in their leisure time they are much freer to experience the city in which some of them choose to settle.
Despite the living conditions, a source of dismay to all migrant women interviewed, most reported satisfaction with their work, which allowed them to save for weddings and contribute to household income, as well as making friends from other districts. Local labor laws worked to prevent excessive overtime (unlike in China), and little discontent with working conditions was expressed. A few reported an increased say in household spending, though these were mostly local women living with their families, rather than migrants, who remitted almost their entire earnings to parents. Parents had typically also played a key role in migration decisions: all interviewees had migrated because of economic distress in their natal families (caused by, for example, illness or death of a parent, natural disaster, or heavy debts). Whether this will lead to more widespread lifecycle migration of young women for accumulation rather than from distress, as in China, is unclear. Several women expressed the wish to continue working after marriage, but in non-factory jobs closer to home. None saw long-term settlement in the city as possible. Significant shifts in migration and urban settlement of young rural women therefore seem unlikely in the short term, given the much lower acceptability of women working outside the home in rural India; and, because women move from one restrictive regime (the patriarchal family) to another (the factory and hostel), emancipatory gains are likely to remain much lower than in China.

Integrating the city: incorporating the villages
A third aspect of both similarity to and difference from the Chinese "model" concerns the experiences of "villages within the city" (chengzhongcun). Unlike city planning, this is not deliberate emulation; nor, as with hiring young female migrants, is it a practice transferred by firms. Instead, the forced incorporation of rural dwellings into the city arises as an unintentional consequence of following the Chinese pattern of large city-style SEZs, and the implications of this for local populations. Despite different land ownership contexts and the lack of any equivalent of China's household registration (hukou) system, the outcomes are in some ways surprisingly similar for former rural dwellers inhabiting the new city. This example therefore highlights how reproduction of a model may happen in unintended ways, including replication (and mutation) of unwanted or unobserved side-effects.
The process of creating urban villages in China and in the Indian industrial city shows strong similarities. In China, urban land is state-owned while villagers own rural land collectively. Only municipal governments, acting in the "public interest", can requisition rural land and change its ownership, enabling transfer of use rights to private parties. There are two main types of rural land: farmland and housing land, the latter forming the village settlement. Only the former is typically requisitioned, in order to avoid paying high compensation to, and costly relocation of, villagers (L. Zhang, 2011). Large stretches of farmland have therefore been converted into urban land, but village settlements remain, unintegrated into new municipal administrative structures (Tian, 2008). This leads to pockets of collectively-owned rural land within Chinese cities.
Scholars emphasize the decisive role of collective ownership in producing urban villages, and the hukou in underpinning differences between "villagers" and "urbanites" (Tian, 2008), yet a similar phenomenon has emerged in the Indian case study. Some of the land which would become the city had been privately owned, while parts had been assigned to landless farmers, but remained state-owned. Villagers who owned farmland were persuaded to sell to a parastatal organization affiliated with the state government at a fixed purchase price per acre of Rs 250,000 (USD 5,600) for barren and Rs 300,000 (USD 6,800) for cultivated land, with the promise of better-paid employment in SEZ factories. State-assigned land was requisitioned for Rs 250,000 per acre. Ownership was then transferred to the private firm which would develop the city. Village habitations were not acquired; farmers remained in village settlements but no longer worked the land.
This mode of development, similar to early Chinese SEZs like Shenzhen though not the more recent CZ1, enabled the case study city to avoid the extensive protests characterizing other cases of Indian land acquisition (Balakrishnan, 2013;Bedi, 2013;Majumdar & Menezes, 2014;Sampat, 2010;Srinivasulu, 2014). Indeed, it has been lauded as a "model" for future development of large SEZs in India. There was no agitation for new compensation mechanisms, and land was bought outright unlike in Maharashtra's Khed SEZ where local landowners gained shareholder rights over redeveloped land (Balakrishnan, 2019). Few locals were resettled, and those who were -for example, for the siting of roads -were provided with single-storey concrete dwellings close to the demolished homes. Those interviewed seemed satisfied with their new accommodation. Protest in the city therefore arose later than in other Indian SEZs, was quickly contained through local police intervention and threats of dismissal from jobs, and centered not on resettlement or land acquisition but on lack of employment and rising inequality, two unwanted "side effects" of the Chinese "model".
While some villages benefited quickly in terms of paved roads and electric streetlights, those more remote from factory construction did not. As in early Chinese SEZs, "offgrid" urban villages emerged, where roads and other facilities remained the responsibility of the lowest level of rural government -here the gram panchayats, which lacked resources to improve infrastructure. Although a new private school was built in the Indian city, with subsidized fees for those formally employed in the zone, a municipal primary school was demolished to make way for new roads, so that children from nearby villages now had to travel several kilometers over unpaved tracks to attend school. Nor were the promised factory jobs available to all, since most villagers' low level of education meant they were not preferred for assembly line work. As in CZ1, and similar to other Indian cases such as Karnataka's Bidadi Industrial Area (Balakrishnan, 2013) and Andhra Pradesh's Polepally (Srinivasulu, 2014), those recruited were typically on informal contracts with local labor agencies, in maintenance, gardening or cleaning. While some villagers found work outside the city gates, many remained unemployed or dependent on the income of a single family member.
Some villagers continued farming beyond the city borders, but the erection of the internal SEZ boundary with its pass system meant that they lost access to grazing routes and watering spots. Remaining cattle were grazed within villages, sometimes causing damage. As in Shenzhen and other early Chinese SEZs, only those employed within the SEZ proper could cross its boundary, and then only at the start and end of shifts. Unlike in China, there were no wage or employment differences between the SEZ and other parts of the city; yet the boundary meant locals' access to space was radically reshaped and the urban fabric of the city disrupted. In Shenzhen, local authorities eventually opened "agricultural gates" for villagers to cross the SEZ to sell produce, and more recently abolished the boundary altogether (Ma and Blackwell, 2017). In the Indian case, similar measures appear unlikely given the SEZ Act's national regulation of SEZ checkpoints, and the current lack of residential or retail areas within the SEZ.
Some former farmers used the cash payment for the land to upgrade their homes, and in wealthier and better located (typically higher caste) villages, rooms were rented to white-collar in-migrants, providing a useful source of income to villagers. This letting of rooms parallels the widespread construction of accommodation for migrant workers in Chinese chengzhongcun, facilitated by villages' exclusion from urban planning regulations. In China this has been so lucrative that some original inhabitants have purchased property in the "city proper", while continuing to let out blocks of apartments in the village (Liu et al., 2010). Cowan (2018) describes similar Indian "elite rentiers" in his study of Gurgaon. However, in the case study city, the number of migrants looking to rent rooms in villages was still low. Unlike in Chinese SEZs, where many migrants operate restaurants or street stalls, the slower pace of development means that such businesses are not yet present in the Indian industrial city. Incoming white-collar employees remain few (and many commute from elsewhere, given the city's lack of amenities), while blue-collar workers reside in hostels outside the city. Moreover, the option of expanding homes for rental was not feasible financially or practically for most villagers, particularly in less well-connected villages without access to paved roads or public transport.
The availability of factory work to only some of the villagers (typically young women educated to at least age 16), and the extension of infrastructural development to only some of the villages, has resulted in prosperity and poverty side-by-side. Whereas in Chinese urban villages, the better-off residents are the original inhabitants, with migrants on lower incomes, in the Indian city the reverse is the case. Villagers who gave up fields lament working as low-paid laborers on the site of land they once owned, while contact with better-paid migrants (and return migrants previously working in nearby cities), and their own fellow villagers now occupying higher status positions, has led to resentment over these new forms of inequality. This forms a contrast to the Chinese context, where discontent has centered on municipal government attempts to redevelop villages, depriving locals of revenue-generating opportunities (L. Zhang, 2011). In the Indian case, zone operators have not (yet) shown interest in redeveloping villages. This may change if the zone is successful in becoming a truly mixed industrial/urban residential and commercial area in the Chinese style.

Conclusion
Through a grounded empirical investigation, this paper has revealed some of the complexities of attempting to import a Chinese "model" of SEZ development elsewhere, here to an Indian industrial city. As the policy mobilities literature suggests, policies cannot be straightforwardly replicated, and this article has shown how the amorphous nature of a "model", with its range of components and actors, exacerbates the challenge. It has highlighted that models are not rigid templates to be copied, but are by nature messy, dynamic and contingent. It is insufficient to represent model mobility merely as intentional adoption by state policymakers and advisors; instead, the roles of non-state actors, including through informal and ad-hoc mechanisms, are crucial. These include deliberate attempted emulation by private SEZ founders, importing of practices by firms, and unintended reproduction (of perhaps unobserved aspects) through interactions with local populations. The process of translating and transfiguring the "model" thus takes place at many different levels and via different mechanisms, including both top-down and bottom-up, intentional and unintentional, strategic and piecemeal.
Our emphasis on non-state actors does not imply that the state is insignificant. National-level policy adoption and adaption -such as the 2005 SEZ Act's mixing of Chinese-style "zoning technologies" (Ong, 2004) with older Indian forms of industrial township and EPZ -provide the essential frameworks for local-level, corporate, and/or individual actors to become agents of policy mobilization. Indeed, the role of the state can be crucial in determining the outcomes of such mobilization: differences in the speed of development of Chinese and Indian zones, for example, may be ascribed both to the difference between state-and private-ownership, and to broader differences in urban development between authoritarian and democratic polities. What is critical here, then, is not that China's SEZs were "exceptional", in the sense of lacking integration with the broader state territory, but that they were experiments undertaken by and for the partystate. Nor is "exception" the most useful tool for examining the adaption of the Chinese "model" in India: though the absence of elected urban governments suggests a distinct governance regime, India's SEZs remain rooted in older, broader patterns of zonal development and private-sector-led urbanization. The Chinese "model" thus interacts with specific local contexts -including differing roles of state and private capital, local and national institutional frameworks, and sociocultural norms and expectations -to create varied impacts. While many studies in the policy mobilities tradition concentrate on tracing how and where policies move and how models mutate (McCann & Ward, 2012), we have shown this focus on policies and expertise to be insufficient without parallel consideration of the context-specific outcomes for the lived experiences of those affected. We argue that this expanded focus is important to understand fully the implications for urban development, social cohesion and individual livelihoods that result from the selective, complex, multi-level adaptation of a Chinese "model" and its interaction with local contexts.
For urban and economic development, the broad consequence has been the Indian zone's achievement of some industrialization and urbanization (falling short, however, of the rapid urban expansion of many Chinese zones). For social cohesion and individual livelihoods, the growth of SEZs in both China and India has raised the question of how multinational capital, and rural and migrant populations, can be integrated into the national urban system, and what sort of urban politics, spatial features, social relations and governance structures can emerge through management of the resulting diversity. In both countries, SEZs have become spaces for classifying and managing populations, as well as producing goods and attracting investments. Both urban villages and migrant women have thus become examples of Ong's (2004) "graduated sovereignty", in which populations are governed differently depending on their relationships to global markets. Yet the implications of this for daily lives and livelihoods in India have, in many and perhaps unforeseen ways, been different from those in China, as key actors, practices and policies remain rooted in local contexts and resistant to convergence. In short, the outcomes of attempting to reproduce a Chinese -or any other -"model" will be complex, different and sometimes unpredicted. They deserve to be better understood through future research. Notes 1. We use the term SEZ in this paper to denote a broad range of economic zones aiming to attract international investment, focus on export-oriented manufacturing and offer incentives for investors. These zones may be officially denoted as SEZs or may include a formal SEZ area as well as other areas. 2. It should be noted that such concerns do not relate only to urban areas: several scholars have highlighted similar land struggles at the agrarian-urban interface (Balakrishnan, 2019;Ghani et al., 2012;Gururani et al., 2021), and the role of newer "transport corridors" as well as SEZs in "liberating" large land parcels for public-private real estate development (Mitra, 2018). 3. CZ1 has not been formally designated an SEZ, but is part of China's newer expansion of zones nationwide with differing nomenclatures from the initial SEZs but similar aims.