I thank the three anonymous reviewers, Professor Charlotte Guénard and Professor Laurence Roudart for their critical comments, advice, and support with accessing the literature which helped in substantively revising the paper.
1The Sriperumbudur-Oragadam region, in the south-western periphery of the Chennai metropolis, in the South Indian state of Tamil Nadu, is projected as India’s Shenzhen. The region is located along the highways connecting Chennai with the major cities of Tamil Nadu and with those in the neighbouring states of Karnataka and Andhra Pradesh. The Government of Tamil Nadu (GoTN) rolled out projects for setting up Export Processing Zones (EPZs)/Special Economic Zones (SEZs), and Industrial Corridors (ICs) to develop the region as a global manufacturing hub for automobiles and electronics (Ernst and Young, 2014; Homm and Bohle, 2012). Vast tracts of agricultural land were notified for acquisition for the implementation of various projects. The sale of agricultural land increased and land prices soared with the announcement of these projects. Private developers announced new projects for the construction of gated housing complexes, malls, and luxury apartments (Raman, 2016). The phenomenon observed in Sriperumbudur-Oragadam is emblematic of the transformation of agricultural land in the villages on the outskirts of the Chennai metropolis. Several scholars suggest that contemporary processes of land transformation result in agrarian households being dispossessed of their land and livelihoods (Levien, 2011, 2018; Vijayabhaskar and Menon, 2018). This paper explores the micro-politics of land dispossession with a specific focus on the responses of affected households to the phenomenon.
2Drawing on qualitative research conducted in Sriperumbudur town and the surrounding villages, the article argues that attending to everyday practices over a period of time can enhance the understanding of the anti-dispossession politics. Tracing agrarian households’ responses to the state-led land acquisition process, the article highlights the limitations of the two dominant theoretical frameworks (Levien, 2011, 2015, 2018; Goldman, 2011) when it comes to capturing the complexity and nuances of anti-dispossession politics. I map agrarian households’ diverse and flexible actions to subvert the land acquisition process, and the differentiated outcomes of their actions. The rambling trajectory of anti-dispossession politics is shaped by the reflexive actions of affected households, by the multiple institutions of the state, and by the socio-spatial networks embedded in everyday relations. Further, the state’s intervention catalysed the forced sale of agricultural land in and around the affected villages resulting in agrarian households, predominantly from the socially and economically disadvantaged Dalit caste, losing their land and livelihoods, thereby reinforcing already existing caste- and class-based inequalities.
3Following Vijayabhaskar and Menon (2018), I suggest that Levien’s (ibid.) definition of land dispossession as a phenomenon solely driven by the state may be simplistic in the Indian context. As shown in this paper, land dispossession in the Sriperumbudur-Oragadam region encompasses the forced loss of land and livelihoods, due both to state-led acquisition and the market process. Levien (ibid.) assumes that agrarian landowners’ participation in the market is always voluntary. By contrast, agrarian households affected by the implementation of SEZs and other types of urban development projects were forced to sell their land for a variety of reasons, including the loss of right of access, or the fear of negotiating with the state over land acquisition. Moreover, state agencies structure the land market through their policies and schemes to facilitate the entry of private developers under private-public partnerships (Baka, 2013; Vijayabhaskar and Menon, 2017).
4In contrast to Goldman (2011), I suggest that landholders’ engagement in the market process is not always driven by speculative intent. Nor do households affected by the land acquisition process benefit from the compensation regime or the market process due to the diverse tenure forms under which they hold land. Both Levien’s and Goldman’s theories focus on households holding land through individual ownership tenure. Finally, both theories (ibid.) overlook the diverse forms of anti-dispossession actions, the role of multiple institutions, and the complex articulation of local and extra-local forces in shaping the rambling trajectory of dispossession politics.
5The rest of the paper is organized into four sections. The next section provides a brief summary of the literature followed (Section 2) by a description of the research location and methods. The findings from the field research on agrarian households’ responses to the state-led land acquisition process, and the outcomes are discussed in Section 3. The concluding section discusses the implications of the findings.
6Two theories are dominantly used to explain the phenomenon of land dispossession that accompanies urban development projects in India, viz. “Accumulation by Dispossession (AbD)” (Levien, 2011, 2012, 2013, 2015, 2018) and Speculative Urbanism (SU) (Goldman, 2011). Levien (2011, 2012, 2013, 2015, 2018) reconstructs Harvey’s (2003, 2007) theory on Accumulation by Dispossession (AbD) to comprehend the politics of SEZs. Drawing on the seminal analysis of primitive accumulation (Marx, 1976), Harvey argues that AbD is an economic process in which dispossession occurs through the transfer of diverse types of assets, viz. land, natural commons, and labour at a very low cost from the vulnerable class to the richer class (Harvey, 2007). The transfer of land and natural commons to globally connected companies under SEZ projects results in dispossessing the rural and urban population of their land and livelihoods (Bannerjee-Guha, 2010; Sampat, 2008; Jenkins, 2007; Levien, 2011, 2015).
7Levien (2011, 2012, 2015, 2018) argues that Harvey’s (ibid.) view of AbD as an extension of the macro-economic process is simplistic, and that it is decisively a political process. Dispossession is facilitated by a “land broker” state, and operates through corporate developers who develop rural land —mainly for technology IT companies and luxury real estate—and profit from the appreciation of land values. Moreover, Harvey (2007) overlooks the anti-dispossession struggles generated by coercive state interventions. The struggles to counter land dispossession share some key features: a focus on a single-issue agenda; overt collective actions; cross-class alliances; autonomy from political parties; and an absence of ideological stance (Levien, 2013, 2015). Actors usually embody different interests. Some are “transactional bargainers” who are interested in higher compensation, and others, “unconditional bargainers” who seek to protect their land.
8In contrast to Levien’s (2013, 2015) scheme of collective action for countering land dispossession, agrarian landowners in Bangalore sold their land to developers (Goldman, 2011). Goldman argues that government agencies using the provisions in law coercively acquire land from agrarian households with little or no compensation, as well as extract high land rents along with large developers. Goldman’s (ibid.) Speculative Urbanism (SU) theory highlights the speculative logic driving the actions of government agencies, large developers, and ordinary citizens in Bangalore.
9Levien does not focus on the reasons for an absence of contestation to the state-led land acquisition process in some places. Other scholars suggest that landowners’ decisions to counter the acquisition process is influenced by a combination of factors including declining agricultural income, lack of space for manoeuvring decisions on land acquisition, the compensation amount, and the availability of alternative livelihoods (Vijayabhaskar, 2010; Chakravorty, 2013).
10Agrarian households in other contexts negotiate with government agencies (Balakrishnan, 2013) or form partnerships with private professionals and developers (Sami, 2012) for land development. Further, affected households try to subvert land acquisition by appropriating the provisions in law and administrative procedures (de Flore, 2015), and mobilizing symbolic and cultural resources (Shivam, 2017). In the light of the above-mentioned evidence, this study explores the responses of affected households to the land acquisition process in the Sriperumbudur-Oragadam region.
11There are two suggestions in the studies reviewed regarding the consequences of anti-dispossession political actions. One is that there is very little or no space for manoeuvring decisions on land acquisition due to the legal and institutional architecture undergirding SEZs and infrastructure projects. In the Indian context, households facing the threat of eviction or land acquisition try to influence state decisions through their elected representatives and local governments (Benjamin, 2000). SEZ projects are housed in parastatal institutions or project-specific special-purpose vehicles to restrict the influence of elected representatives (Benjamin and Bhuvaneswari, 2006; Idiculla, 2016). Moreover, government agencies mobilize specific legal provisions such as eminent domain, administrative processes, or administrative categories to quell the opposition to land acquisition for SEZs and infrastructure projects (Ramanathan, 2011; Baka, 2013, Levien, 2011).
12By contrast, other studies show that decisions on land acquisition are shaped by back and forth negotiations between various scales of the state and the elected representatives (Balakrishnan, 2013; de Flore, 2017). In addition, Chakravorty (2013) argues that the high land prices in India, together with the high compensation provisions under the new Land Acquisition, Rehabilitation and Resettlement Act (LARR) would make land acquisition too expensive and stall urban development projects. Nielsen (2018) critiques Levien’s scheme of land dispossession as a generative event of anti-dispossession politics, explaining that it diverts attention from the broader consequences of land acquisition on everyday life especially on caste, class, and gender relations.
13In light of the competing evidence in the studies discussed above, this paper explores the forms and effects of anti-dispossession politics and the reasons thereof. The related findings are discussed in Section 3.
14Tamil Nadu was the first state to formulate an SEZ policy in 2003, before the Government of India enacted the Act in 2005. The GoTN announced the policy to set up special economic zones in 2000, and established Mahindra City, the country’s first operational SEZ, under a public-private partnership arrangement in 2002. Forty-six of the ninety-nine SEZs approved in 2012 by the Government of India are located in and around the Sriperumbudur-Oragadam region (GoI, 2012). Prior to forming SEZs, the GoTN created EPZs, two of which are located in Sriperumbudur. In 2002, all EPZs were declared as SEZs (Chandrachud and Gajalakshmi, 2014; Sashi Kumar, 2008).
15The State Industries Promotion Corporation of Tamil Nadu (SIPCOT) and the Tamil Nadu Industrial Development Corporation (TIDCO) created the EPZs/SEZs in the Sriperumbudur-Oragadam region. The Government of Tamil Nadu established TIDCO in 1965 and SIPCOT in 1971 for promoting industrial growth. Both agencies were formed under the Companies Act. Besides SEZ, TIDCO partnered with developers to form townships (Raman, 2016). While SIPCOT is a nodal agency to finance and create industrial infrastructure, TIDCO supports the establishment of medium and large industries in association with the private sector. TIDCO is also registered as a non-banking financial institution with the Reserve Bank of India. Until the mid-nineties, SIPCOT acquired land from farmers and allocated developed land to industries. Its policy shifted in 2005 towards public-private partnerships and land-based financing. Both agencies accumulate land for their banks and also facilitate land acquisition for developers through the market process. Until 2015, SIPCOT acquired around 7018 acres in the region (MSME-DI, 2016).
- 1 Official heading the subdistrict offices of the Revenue Administration. Each sub-district or tehsi (...)
16In the Indian context, the Revenue Department (RD)—a regional government agency headed by the District Collector—is the only agency authorized to acquire land. SIPCOT and other agencies put in a request to the RD to acquire the identified land for infrastructure projects. The collector appoints a special Tahsildar1 to acquire land for a specific project. The RD initially notifies the landowners about the decision to acquire their land after obtaining a no-objection certificate from the rural local government, i.e. the village panchayat. The department notifies different categories for land acquisition for SIPCOT projects, which include land under agriculture, common land, habitation inside the village, and residential layouts formed with the permission of the Department of Town and Country Planning (DTCP). The RD’s notification for acquisition stands even if a residential layout may have planning approval. The RD transfers the acquired land to SIPCOT, which in turn transfers it to private companies after infrastructure development, or to developers for land development under the public-private partnership arrangement.
17Findings discussed in this paper draw on research conducted in three phases between 2013 and 2018 in Sriperumbudur town and twelve villages in the region. In the first phase, I studied eleven villages where the RD acquired land between 1997 and 2002 for setting up three EPZs/SEZs in Sriperumbudur, Oragadam, and Irungattukottai. The largest automobile hub is now located in the Oragadam SEZ and a high-tech industrial park, in Irungattukottai. Subsequently, the Revenue Administration (RA) acquired land in other villages of the region in 2005, 2011, and 2015. I revisited five of the eleven villages in 2016-17 and, in addition, visited seven villages which were affected by different projects in order to map the responses to and effects of land acquisition. The field research draws on ethnographic methods. Information was collected through observation, interviews, and conversations. Interviews were conducted with field officials of the Revenue Administration (village administrative officer, Revenue Officer), village panchayat leaders, landholders, brokers and tenants in each village, officials of a cooperative society, and investors to gather basic information on the number of households, caste composition, land area, categories of land, process of land acquisition, responses, history of land sales, characteristics of landowners and tenants, and land conflicts. Besides this, I analysed the court judgements of the land acquisition cases in Sriperumbudur. In addition, a rapid survey of sixty farmers across five villages was conducted in the second phase in order to map the changes in land prices, current source of livelihoods, and shifts in landholding patterns. I also followed a conflict between developers, landowners, and investors to trace the market process. I conducted the interviews in Tamil and maintained field notes in the same language.
18This section illustrates landholders’ diverse responses to the state’s land acquisition process. It describes landholders’ political actions to contest the land acquisition process, followed by the reasons for the absence of contestation in some contexts. Further this section analyses the phenomenon of dispossession through the market process.
19Landholders constituted a heterogeneous group, in terms of the size of their landholdings, caste and class affiliation, and the tenure forms under which they held land. The size of their landholdings varied from an area of less than an acre to a maximum of fifteen acres. Agriculture was the main source of livelihood for the different categories of landholders at the time of land acquisition. Based on the sizes of holdings, landholders can be categorised as large landholders who owned between ten to fifteen acres, medium holders having between five and nine acres, and small and marginal farmers possessing from less than an acre to a maximum of 3 acres.
- 2 The various categories of poramboke land include natham (village land used for habitation), meicha (...)
20The few large landowners were from the Brahmin-caste and the middle-caste communities including Naidus and Naikkars. The Brahmin caste owners were absentee landlords, while those from the middle-caste communities resided in the villages and dominated local politics. The mid-size landholders from different castes, viz. Naikkars, Naidu, and Vanniyar, cultivated the owned or leased land. The marginal farmers and the landless households were predominantly from the Dalit, with few from the middle castes. The Dalits depended on grant land, and various categories of common land or poramboke land2 for their livelihoods.
21The RD issued an order to landowners of the Sriperumbudur–Oragadam region ordering them to surrender their land voluntarily along with the land documents. In addition, the department issued notifications regarding the acquisition of the land inside the village occupied by the Dalit households. Landholders using lease, grant, or poramboke land did not receive any notification.
22During the first round of acquisitions in 1997, resident and non-resident households possessing agricultural land complied with the RA’s order. Those who had invested in a plot in the private residential layouts, henceforth referred to as plot owners, individually filed cases with the Madras High Court contesting the acquisition order.
23In the subsequent round of the land acquisition process between 1998 and 2000, landholders across eleven villages contested the acquisition process, engaging in a range of actions including non-compliance (Razzaz, 1994), quiet individual negotiations via their political and caste networks, lobbying through political party leaders, collective protests, legal action, and the sale of land. Initially, landowners assumed that their non-cooperation would force the RD and SIPCOT to abandon their plan. Few landowners individually attempted to influence the land acquisition officer to exempt their land from the notification order. Facing difficulty in actually acquiring land, RA and SIPCOT officials invited agricultural landowners and plot owners to three rounds of negotiations. Landowners demanded the cancellation of the land acquisition order, but the officials insisted on negotiating only the terms of compensation. The meetings ended in an impasse, as the landowners refused to surrender their land. With the failure of negotiations, the RD issued an ultimatum, threatening to acquire the land using force and without any compensation.
24Realising that losing their land was inevitable, resident landowners from different castes across the eleven villages came together in 1999 to secure their land (and not for compensation). An ex-panchayat chairman from one of the villages, along with a local political activist, organized not only landowners but also others who possessed land under different tenure forms. The group engaged in various forms of collective actions including representations to senior party leaders, public hearings, petitioning and lobbying the District Collector, the land acquisition officer and SIPCOT officials, and protests. The group drew on the support of a network of lawyers in Sriperumbudur town, a Dalit activist organisation, and a non-governmental organisation in Chennai. In the meantime, RA and SIPCOT officials announced an enhanced compensation amount, which was again rejected by the villagers. When their protests and lobbying did not yield much result, they decided to approach the courts.
- 3 However, there is ambiguity as to what constitutes public purpose (Ramanathan, 2011). The new land (...)
- 4 https://indiankanoon.org/doc/1924061/ (Retrieved January 2019).
- 5 See also Haridoss vs Government of Tamilnadu, writ petition number W.P.No.16922 of 2007.
25The District and the Madras High Courts adjudicated two types of cases on land acquisition in the Sriperumbudur-Oragadam region. The first one concerned the cases filed individually by plot owners, who appealed for exempting their land from land acquisition. The government mobilizes the power of “eminent domain” to acquire land for SEZs and other infrastructure projects. Eminent domain is premised on the twin principles of public purpose and just compensation. It allows for the compulsory acquisition of private land for public purposes.3 When land is acquired under the eminent domain article, the courts adjudicate the disputes within the framework of compensation. Some landowners appealed individually for exemption, stating that their land was already developed or that their residential layouts were approved by the planning authority, namely, the Department of Town and Country Planning. The court judgement exempted land parcels under industrial or religious use under the provisions of the Land Acquisition Act of 1894,4 but supported the acquisition of privately developed layouts even though they may have had planning permission5 (Rao, 2010).
- 6 Refer Madras High Court writ petitions numbers filed in 2000: 16922, 29294, 29090, 29113, 29707, 2 (...)
26The second type of case filed by the representatives of agrarian households in the eleven villages were pleas to cancel the acquisition order on the grounds of loss of livelihoods and of the inability for petitioners to find alternative jobs due to a lack of skills.6 The case was heard by the District Court and subsequently by the Madras High Court between 2000 and 2010. In 2005, as the High Court was hearing the case, RA-SIPCOT officials threatened to take the land by force. By then, the Special Economic Zone Act of 2005 and a Tamil Nadu Extraordinary Gazette Order authorizing the use of force for land acquisition had been passed. The court judgement pronounced in 2010 upheld the land acquisition for SIPCOT’s project, but ordered the RA to enhance the compensation amount (Rao, 2010).
27Landowners and government officials often disagreed on what constituted adequate and just compensation. Landowners’ expectations were based on the value of their agricultural land—which depended on the type of soil (wet or dry land), crops, and irrigation facility. The market value for wet and dry agricultural land differs. The RA fixed the compensation amount on the basis of the market value of dry agricultural land for all types of land. Further, the RA took into account the location of the land in relation to the main road. A remote wet land parcel was valued less than dry land closer to the road. There was also a disagreement about the estimation of the market value (Rao, 2010). The RA fixed the market value on the basis of sale prices reported in the registered sale deed, which is often less than the actual market price of land. In their responses to a court query, the Revenue officials stated that the market value of land at the time of acquisition was an inflated value as by then real estate developers had already purchased the land and hiked the price; thus, it could not be used for fixing compensation (Rao, 2010).
28In addition, thirteen Dalit households from one of the villages possessing grant land approached the Madras High Court7 in 2008 to restore their grant land. The court dismissed their cases due to two reasons. The grant land, locally known as Panchami land, was allotted to Dalit households under a land reform movement called the Bhoodan movement. The movement encouraged landowners to gift/donate part of their land to the landless in their respective villages through the Bhoodan Board, a government agency. The Bhoodan Board allotted land to some households, but did not transfer the land documents to the allottee. In other cases, large landowners who pledged to donate their land did not actually transfer it to the board. The inheritors from these families sold the land. Between the mid-sixties and 2001, Panchami land was exchanged three to four times and the sales were registered (Raman, 2016). As these sale transactions happened outside the village, Dalit households possessing Panchami land came to know about it only after SIPCOT allotted it to companies. However, the Madras High Court allowed the affected Dalit families to appeal again to the civil court for compensation. Although the civil court ordered the RA to return the acquired Panchami land to the respective households,8 the revenue officials stated that in practice, it would be difficult to retrieve the land as the companies had enclosed it and their operations had been in full swing since 2006.
29The courts’ intervention closed the possibility of further negotiation through political or administrative channels as it is binding on the state to implement the court order. Large landholders from the Brahmin caste favoured the judgement as they perceived limited chances of reclaiming their land from tenant farmers, given the state’s history of anti-Brahmin caste politics. Both large and mid-size landowners from the Naikkar and Naidu castes decided to settle for the enhanced compensation amount, fearing the possibility of losing both land and money. Small and marginal farmers from the middle-caste and Dalit-caste communities resented the decision of large and mid-level farmers as they did not benefit from the compensation regime, and it resulted in the disintegration of their collective action. Small and marginal farmers from the middle-caste communities who surrendered less than three acres stated that the compensation amount was inadequate to invest in an alternative livelihood. Although landowners tried to re-negotiate the compensation norms, they were not successful as by then the collective action had dissipated. In addition, developers and investors privately developed residential layouts held substantial portions of vacant land at the time of the land acquisition, which was acknowledged by the land acquisition officers in their responses to court cases on land acquisition (Rao, 2010).
30Overt collective action was absent in the subsequent rounds of land acquisition for SIPCOT’s projects in 2005, 2008, 2011, and the most recent one in 2016. The differentiated outcomes of the court case and collective action in the eleven villages influenced some landholders not to contest the land acquisition process. Further, households possessing “poramboke” land or grant land (Panchami land) did not resist acquisition as they perceived very little chance of any redressal. In their view, not only are legal disputes costly and time consuming but also the chances of regaining possession of their land is difficult even after a favourable court judgement. Interviewees whose land was acquired stated that securing the compensation amount from the RA and SIPCOT was a time-consuming process and that it took them two years of regular visits to the offices and moving the files between different desks to secure their compensation order. Even though the court judgement for enhanced compensation was passed in 2010, several medium landowners did not receive the compensation amount until October 2017.
31Findings reported in the above sections highlight the limitation of Levien’s (2011, 2013, 2015, 2018) scheme of anti-dispossession politics. Not all landowners contested the state’s land acquisition process. Those who contested the acquisition process engaged in various forms of individual and collective actions for which they flexibly drew on their alliances with political parties, caste networks, professionals, and institutions. Participants’ interpretation of the conflict at different times influenced their engagement in various forms of action. The affected households negotiated the land acquisition process engaging with multiple institutions. The use of eminent domain, the location of the projects in parastatals, the administrative categories of land, and the entries in official records (which often do not reflect possession on the ground as also reported by de Flore, 2017) restricted the political and legal space that the affected households had for negotiating the land acquisition process. The role of the courts is critical in mediating conflicts over land acquisition. The reflexive actions of different actors and institutions influenced the rambling trajectory of the conflict.
32Second, while compensation is certainly a factor influencing the emergence of anti-dispossession politics as some scholars suggest (Vijayabhaskar, 2010; Chakravorty, 2013), it is not the only one. A combination of other factors influenced the affected households’ tendency to contest the land acquisition process, such as perceptions of space for manoeuvring government decisions, the experience of their networks with anti-dispossession politics, the capture of land by developers, and the land tenure forms.
33Third, landholders’ actions to contest the acquisition process resulted in differentiated outcomes reinforcing caste- and class-based inequalities about which there is little in Levien’s (2013, 2015) scheme. Dalit households did not benefit from the compensation regime as they depended on poramboke (common) land, leased land, or grant land for their agriculture.
- 9 Based on a survey of 60 farmers in 4 villages (Oragadam, Chettipalayam, Vadakal, and Senthamangala (...)
- 10 1 USD = 73 Indian Rupees (Rs).
- 11 1 cent = 40.47 sq.metre.
34The announcement of SIPCOT’s SEZ projects accelerated the sale of agricultural land outside the village habitation. Large and medium landowners in the villages affected by land acquisition and in the neighbouring villages sold their land. Landowners sold their land due to a combination of factors such as: the loss of right of access to their land and irrigation facility with the acquisition or sale of neighbouring land; the enclosure of common grazing land and wetlands; the fear of acquisition by the state; and the deterioration of irrigation facilities. According to the village administrative officer of Oragadam and the brokers in the region, land sales had peaked since 1998 in different villages just before the RA issued the notification for land acquisition. Although the sale of agricultural land and its conversion for urban use existed in the seventies and eighties, it involved a smaller area, between one acre and three acres at a time (Raman, 2016). The area of land transacted increased with the entry of corporate developers (Raman, ibid.). The sale prices9 reported by landowners suggest that the price of land increased fivefold between 1998 and 2005, from around 13.69 USD10 per cent11 to 68.50 USD per cent, and doubled over the next two years, between 2006 and 2008, from 68.50 to 82.91 USD. Land prices quoted in 2012 were around 164 USD per cent for wetland and 109 USD per cent for dry land.
35With the exception of some large to mid-level landowners from the Brahmin-caste and the middle-caste communities, other medium and small farmers depend on their sale proceeds or compensation amount for their household expenses. Large landowners from both the Brahmin-caste and the middle-caste communities ploughed the proceeds from their sale or compensation amount in land, or built property for future gains. In addition, landowners from the middle-caste communities benefitted from the growth of rental markets for housing, transport, and construction equipment. They upgraded their residences in the village and constructed new apartments outside the village for renting to different classes of migrant labour. Besides this, they purchased construction equipment and automobiles, which they rented to the multinational companies.
36Medium and small landowners from middle-caste communities predominantly used their sale proceeds and compensation amount to meet lump sum expenditures such as marriage or children’s education, and to upgrade their houses for renting on a small scale. Those who could not invest in upgrading their houses leased out land for small businesses. They depend on their rental income together with wage employment and meagre interests from bank investments to meet their household expenses. Households which once owned 5 acres or more of agricultural land in 2000 held an acre or less in 2017. Those who sold less than an acre lost their land and depended on daily wages. The rental market slumped with the closure of Nokia and two other factories at one of the three SEZs. Apart from property owners, the downturn affected small retail businesses.
37The tenants are predominantly migrant labourers from different Indian states including Tamil Nadu, Bihar, Jharkhand, and North-eastern states of India. Many of them work as contract labourers in the companies at Sriperumbudur-Oragadam SEZs. They find their jobs and accommodation through labour contractors. Both male and female migrants are contract labourers and move to the city alone, and between six to eight individuals share one-room accommodation. The rental values are 21 USD per month for a room in a thatched hut; 27 USD to 34 USD for one room with tile roofing; and between 41 USD and 68.50 USD per month for a one-bedroom house with a reinforced cement concrete roof. In general, there is hostility towards inter-state migrants whom the villagers blame for rising crime in their area.
38The socially and economically weaker Dalit households across the different villages lost their grant land as well as the right of access to common land. They are predominantly employed as daily wagers in Sriperumbudur town or in Chennai. They do not have the capital to invest in house construction for renting and continue to live in houses with either thatch or asbestos sheet roofing. At the time of land acquisition, SIPCOT promised jobs in the new companies to a number of the affected landowners. Youth predominantly from the Naidu and Vanniyar caste households, who had completed their college education, were offered contract labour for a year on the factory floor.
39Landowners sold their land to a heterogeneous group of developers, brokers, and investors (Raman, 2016). While developers and investors with varying scales of operation benefitted from speculation, the scale of speculation differed. Large developers operating at the national and global scales and large investors benefitted the most. They control large contiguous tracts of erstwhile agricultural land outside the habitation area of the villages. Large investors purchase land in the region expecting assured high returns. Both large developers and large investors prefer to purchase already assembled land from small and medium developers to minimize the risks involved with negotiating with several landowners for assembling a large tract of contiguous land. Small and medium developers assembled land incrementally over the years by mobilizing their caste and political networks. They sold the assembled land to large developers as their projects were affected by SIPCOT’s land acquisition process and a 2015 court judgement prevented the sale registration of plots in privately developed residential layouts.
40Land market transactions are not always smooth, and conflicts between developers, investors, and landowners are common. Developers with weak local ties face the risk of landowners withdrawing their sale. A dominant practice among developers and landowners is to enter into an unregistered sale agreement and to pay the sale amount in instalments. Developers obtain a general power of attorney from the farmers to parcel the land for sale to retail investors. They use the amount collected from buyers partly to pay the landowners and partly to develop the land or invest in more land in the region. A delay in the project cycle can result in retail investors holding back their payment, which in turn affects the developers’ payment to landowners. Such delays in payment can extend for several years, by which time land prices have escalated. In some villages, landowners collectively demanded a higher price for their land and refused to register the sale deed. With the support of local party leaders, the landowners collectively reoccupied their land and cultivated it. There are several instances of developers not paying the small farmers, or projects being abandoned midway, affecting retail investors. These conflicts between landowners, developers, and investors are neither easily audible nor visible, and play out in different politico-administrative spaces, viz. the local government (or panchayat) and the arms of the Regional Government namely the RD, the Department of Housing, and the Department of Town and Country Planning (DTCP).
41The findings summarised in this section call for a nuanced reading of the land market dynamics. The speculative urbanism theory tends to homogenize actors, their interests, and their ability to tap into speculative markets. As can be inferred from the above discussion, the market process has led to differentiated outcomes among landowners, developers, and investors. It has reinforced the already existing caste and class inequalities within the villages and introduced new inequalities—between the old and new settlers. Large landowners from the higher castes and the middle castes, large developers, and large investors profited from the speculative markets. A majority of medium and small landowners were forced to sell their land. They moved into unstable and low-paying jobs after losing their land either completely or partly. Similarly, retail investors in such locations include households from different income groups aspiring to own a house whose investment in land is not only driven by speculative gains. Following Vijayabhaskar and Menon (2018, 2017), I argue that it is important to differentiate between the reasons for engagement in the market process and the circuits of investments in land for a nuanced understanding of the consequences of markets on different actors.
42Moreover, the bargaining power of different scales of developers vis-à-vis one another and between a developer and an investor is neither straightforward, nor stable over time. Land sales slowed down post 2012 and further declined post the floods in 2015. Several large developers’ projects were incomplete and the units built were unsold. These projects were financed through credits from local banks and capital markets, and loans to real estate developers pose a risk to the financial institutions (see also JLL, 2014). When I revisited the villages in 2018, large landowners were waiting for potential buyers so that they could sell their land. Although the quoted value of land was as high as 50,000 to 80,000 USD per acre, investors were reluctant to purchase land. While the court interventions curtailed the actions of small and medium developers in 2015, another judgement in 2017 removing the ban on registration opened up space for the formation of small- to medium-scale private residential layouts. The Real Estate Regulation Act was also introduced in 2017 and its impact on the market process is not yet known.
43This paper is an attempt to broaden the discussion on the phenomenon of land dispossession beyond a single moment of state-led land acquisition process and overt political actions.
44The findings summarised in Section 3 illustrate the limitations of the two dominant theoretical frameworks (Levien, 2013, 2015; Goldman, 2011) when it comes to capturing the complexity and fluidity of anti-dispossession politics. Firstly, in contrast to Levien’s suggestion, an event of land acquisition does not always catalyse resistance actions. Secondly, while Levien’s (2015) scheme emphasises overt collective action, the SU theory overlooks any resistance action to the land acquisition process. Collective action is one of the various forms of action enacted by the affected households to contest land acquisition. The role of networks embedded in everyday relations of caste and place is a factor influencing the mobilization for collective action.
45Agrarian households in places like Sriperumbudur-Oragadam are forced to sell their land, but their participation in the market is not always driven by speculative intent. Landowners, developers, and investors constitute a heterogeneous group and their ability to benefit from speculation is not uniform. The SU theory overlooks the diversity of interests of different players and their reasons for engagement in the market process. Following Vijayabhaskar and Menon (2018), I argue that it is important to differentiate various actors’ interests, the ability to speculate, and the actual gains from the land market.
46Both theories do not take into account the role of multiple institutions shaping the outcomes of the conflict. The politics of land dispossession is not a linear contest between a government agency and the affected households. Through a focus on the process, the paper has shown how the trajectory of the conflict was shaped by multiple institutions and external forces. As noted by other scholars (de Flore, 2017 ; Balakrishnan, 2013), the discursive trajectory of anti-dispossession politics is reflexively influenced by the back-and-forth negotiations between the affected households, government agencies, and the courts. The interventions of different institutions not only introduced unexpected turns, but also reflexively influenced both the government institutions and the agrarian households to adjust their demands as the conflict unfolded. The political-economic shifts such as the slump in land markets, economic downturn, and courts’ interventions can alter the bargaining power and fortunes of the different actors, as shown in this paper. I also highlighted the differentiated outcomes of their political actions. Further, in contrast to the findings of other scholars (Chakravorty, 2013; Vijayabhaskar, 2010), I have shown that high compensation and the availability of job opportunities are not the only factors for the absence of resistance action.
47Caste as an analytical variable is limited in the existing theories. The compensation regime benefits largely developers and large farmers from upper-caste and middle-caste communities. They also benefit from the new local economic opportunities in the forms of rental markets, real estate in land, and retail trade. The socially and economically disadvantaged Dalits and economically less well-off middle-caste households are adversely affected.
48Land is not an inert object over which politics play out. Rather, the materiality of land influences the forms and outcomes of anti-dispossession political actions. The fluidity of land politics calls for a nuanced reading, as the process is influenced by a variety of forces including market shifts, conventions, and the state’s actions. A focus on everyday practices over a prolonged period is important for capturing the nuances and fluidity of this politics.