Exceptional growth at an unexpected location : what is driving India's rural-urban transition? : contrasting theories in a case study in Bihar

(1)

E

XCEPTIONAL

G

ROWTH AT AN

U

NEXPECTED

L

OCATION

:

W

HAT IS DRIVING

I

NDIA

’

S

R

URAL

-U

RBAN

T

RANSITION

?

Contrasting theories in a case study in Bihar

Steven de Jong 10266925 jongstevende@gmail.com Master’s Thesis 31 December 2018 Research Master Urban Studies, Graduate School of Social Sciences, University of Amsterdam Supervisors: Prof. dr. Karin Pfeffer and Robbin Jan van Duijne MSc (PhD candidate) Second reader: dr. Fenne Pinkster

(2)

2

Cover picture taken by author on 22 February 2018.

(3)

3

Acknowledgements

First of all, I owe a debt of gratitude to the many people that participated in this study in one way or another, taking the time to share personal stories without any self-interest, and usually with the greatest enthusiasm. Furthermore, I am most grateful for having encountered both Mrityuntjay Jha and Bhushan Kumar, who proved to be indispensable for conducting my research on the ground. Vishnuram Abhinav and his parents supported my research both practically in Bihar as well as with useful ideas at IIT Bombay.

I could not have wished for a more motivated team of supervisors than Karin Pfeffer and Robbin Jan van Duijne. With the time, energy and patience spent over the last two years they easily exceeded the official requirements for supervision. Their priceless motivation helped me through the difficult moments on the road and got me interested in this research field in the first place. Their feedback always found the weakest spot in my arguments, bringing my work to a higher level, and I genuinely enjoyed our discussions, especially when we did not agree.

My friends Julia Jansen, Jonas Kertscher and Lotte Frissen made staying in Mumbai feel like having a family around. I will never forget discovering India and its cities with you, as well as exchanging our urban ideas – whether related to our research or not. I am further grateful to professor Arnab Jana and Nikhil Kumar at IIT Bombay and Kshitij Batra at IDFC Institute for providing useful insights during our lengthy discussions.

Lastly, I will name some of the people who helped me in various ways while creating this thesis. Cecile Hooijdonk, thank you for always asking, listening, and offering welcome distractions. Luuc Brans and Isabelle van Elzakker, your scientific insights helped me during the writing and you proved that I can always count on you, even with 7,000km and a bad internet connection in between. And Floris van Dort, our numerous coffee and sandwich breaks were not just necessary but more than a pleasure – I am grateful that we got to start and finish our time as a student together.

(4)

4

Abstract

With worldwide urbanisation reaching exceptional levels and economic activity increasingly clustered in urban environments, India seems to deviate: it is witnessing swift economic growth despite being one of the least urbanised countries. Indian megacities are usually the centre of attention, while urbanisation is occurring increasingly in small towns in the rural-urban transition. Little is known about this transitional zone. In particular, the dispersed pattern of rapidly urbanising settlements organised around a nonfarm economy, located in the rural, poor and populous state of Bihar, is opposing our prevalent understandings focused on agglomerations and cluster effects. Therefore, this case study investigates the drivers of growth in Bihar’s rural-urban transition in order to determine the suitability of these theories. An inductive approach, with 95 interviews as the main data source, serves to answer the following research question: “how does the economic geography evolve at the rural-urban transition in Bihar, and how does this differ from conventional understandings focused on a central agglomeration?”. I find that the local economy of settlements in this area typically consists of numerous small, consumption-focused, self-employed businesses. The perception of future prospects, influencing migration patterns, is seen as key to understand urbanisation in this area. People find themselves pushed away from the countryside due to precarious livelihoods while living conditions in large agglomerations are deteriorating. At the same time, the presence of schools and the perception of better livelihoods in the urban economy is making middle-sized, nearby settlements increasingly attractive. I argue that the explanatory power of agglomeration-focused theories diminishes for the rural-urban transition, and instead, a hybrid of push and pull drivers specific for the area should be adopted.

Key words: rural-urban transition, subaltern urbanisation, Global South, India, Bihar, livelihoods strategy, nonfarm economy

List of Abbreviations and Acronyms

CT Census Town

GDP Gross Domestic Product

IIT Indian Institute of Technology

NEG New Economic Geography

NWO Netherlands Organisation for Scientific Research

OECD Organisation for Economic Cooperation and Development

RUTZ Rural-Urban Transition Zone

SGDP State Gross Domestic Product

ST Statutory Town

UvA University of Amsterdam

(5)

5

List of Figures

Figure 2.1: McGee’s spatial configuration of an Asian region 15 Figure 4.1: bridge over the Burhi Gandak river, Samastipur 31 Figure 4.2: Map showing Samastipur’s location 33 Box 1: Benchmark profile of agglomerative growth: Muzaffarpur 35 Figure 4.3: Large new structures built in rural area between Muzaffarpur and the highway 36 Figure 4.4: Shopping mall with cinema in Muzaffarpur 37 Figure 4.5: Map with distribution of cities and towns in Bihar 39 Figure 4.6: Map of villages and census towns with urban characteristics, adjacent to Samastipur 40 Figure 4.7: Settlements containing built-up area adjacent to Samastipur ST 44 Figure 4.8: New bridge over railway lines 45 Figure 4.9: Congestion on river bridges 46 Figure 5.1: Rural area just outside Samastipur 47 Figure 5.2: Weekly farmers’ market 49 Figure 5.3: Weekly farmers’ market 50 Figure 5.4: Small agricultural land plots 52 Figure 5.5: A plot of farmers’ land surrounded by built-up area in Adarsh Nagar, Samastipur 53 Figure 5.6: Rural settlements enclosed by new urban built-up area in Adarsh Nagar, Samastipur 53 Box 2: Profile of a farmer 56 Figure 6.1: Bridge over railway line 59 Figure 6.2: Location of respondents in Samastipur 61 Figure 6.3: Bangali Toli main street, Samastipur, 10 a.m. 66 Figure 6.4: Bangali Toli main street, Samastipur, 13 p.m. 67 Figure 6.5: New electronics store in Gudru Bazar, Samastipur 67 Figure 6.6: Respondent 34, a typical footpath stall at location type 2 69 Figure 6.7: Variety of products and sellers on a part of Bazar Samiti regional market, Samastipur 70 Figure 6.8: Consumption-focused salesmen on the countryside (near Khanpur) 71 Box 3: Profile of a small business owner 73 Figure 7.1: Bicycles parked in front of class rooms at an intermediate school. 77 Figure 7.2: Classroom at Genius Commerce School 79 Box 4: Profile of a student 79 Figure 7.3: Meeting for new business owners 84

List of Tables

Table 3.1: Participants of semi-structured and structured interviews 26 Table 3.2: Respondents of in-depth interviews 27 Table 5.1: Basic characteristics of respondents at weekly market 49 Table 6.1: Respondents in Samastipur 60 Table 6.2: Change in population and share of nonfarm workers in Samastipur, 2001-2011 62 Table 6.3: Classification and location of economic activity of respondents 64 Table 6.4: Regional function of Bazar Samiti shops 65 Table 7.1: Characteristics of respondents (students) 79 Table 7.2: Future prospect of respondents 82

(8)

8

CHAPTER 1

Introduction

“In ten years, this area will be completely urbanised. Modernity comes in all its aspects. And I will have a better future than my father, who was a farmer”1_{. This statement of an Indian}

resident on the edge of a rural settlement reflects how many foresee an urban future, and the current trend of unprecedented urbanisation all over the globe has received ample attention in both science and society. The projection of a worldwide urban population of 6 billion by 2050 – a sixfold increase since 1950 – is often used to illustrate the relevance of studying urban growth (OECD, 2015; United Nations, 2014). City life is regarded as providing its residents better opportunities, and urbanisation is seen as a trend that allows us to improve living conditions for many (ibidem). Now that half of the world’s population lives in a city, it is for some even reason to speak of an 'Urban Age' with cities as the centre of attention of scholars and policy makers alike (Brenner & Schmid, 2014).

Despite Western countries having become almost completely urbanised and most urban research focused here, Africa and Asia’s major agglomerations are nowadays the main location of urban growth. The significant number of slum dwellers in these cities is seen as a threat to sustainable urban development and an opportunity to improve living conditions for the world’s poor at the same time (United Nations, 2014, pp. 2, 37). Turning to one of the world’s most remarkable cases, India appears to deviate from the Urban Age these with only 31.2% of its population living in cities. However, change is occurring rapidly and at a large scale: the urbanisation rate increased with 3.4 percent point between 2001 and 2011, adding 90 million Indians to the urban population in a mere ten years (Government of India, 2001, 2011). Many of India’s urbanities live in megacities – 46 cities in the country have an official population of over a million (ibid.) – and most policy and science is focused on these cities accordingly. But again, a shift is occurring in the locus of urbanisation, this time from megacities to smaller towns located at the less developed rural-urban transition (Denis & Marius-Gnanou, 2011; Pradhan, 2013; Schoenberger & Walker, 2016).

There are more reasons to shift the focus of urban research to India’s countryside. While the concomitant growth of industries and urban centres in Western countries is uncontested, various countries in the Global South – and India in particular – are witnessing a contrasting reality (Turok & McGranahan, 2013). The fact that India’s swift economic growth currently occurs in a seemingly rural environment challenges the conventional belief that processes of urbanisation and economic growth are mutually reinforcing and moving in tandem2_{, focusing on agglomeration economies and spatial concentration of people for}

explaining urbanisation (Government of India, 2011; Nijman, 2012; Scott & Storper, 2003;

1_{Source: interview 52, business owner in Musapur, Samastipur.}

2_{A large body of literature sustains the relationship between urban growth and development and modernisation. See among others}

(9)

9

Von Thunen, 1966). Although indications are found, not enough research is done in order to establish that the case of India actually represents a deviant causal path. Further knowledge gaps exist regarding the specific dynamics and conditions under which the linkages between urban and economic growth operate, and this is particularly ill-understood in the Global South (Overman & Venables, 2010; Turok & McGranahan, 2013).

One of India's most extraordinary cases is found in the populous northern state of Bihar, where urbanisation at first glance appears to be almost decoupled from economic growth. Urbanisation, with a rate of 10.8% in 2011 according to the census definition, is extremely low and displays a dispersed pattern with an exceptional share of Bihar’s urban growth occurring in small and poorly connected settlements (Government of India, 2017a). At the same time, Bihar’s GDP of just 450 euros per capita3_{makes it economically one of}

India's weakest yet most rapidly developing states (Central Statistical Organisation, 2016). Small, nonfarm economic enterprises appear to be the main contributor to growth, in disparity with the agricultural identity of the region (Government of Bihar, 2018).

The rural and poor yet densely populated rural-urban transition zone (RUTZ) in Bihar is at least challenging, if not contrasting prevalent theories of urban growth that link economic growth to large agglomerations. Alternative theories established in the Global South shift the origin of growth from large agglomerative forces to a range of endogenous factors and emphasise the local push and pull dynamics of migration (Mukhopadhyay, Zerah, & Denis, 2017; Ravetz, Fertner, & Sick Nielsen, 2013; among others). In aiming to first explore the drivers of urbanisation patterns at the urban-rural transition and secondly investigate whether this provides a deviant case to accepted conceptualisations of urban growth, Bihar is the case of this study. I have selected the middle-sized settlement of Samastipur, located in Bihar’s countryside, to allow for an analysis of the relationship between economic growth and urbanisation in the RUTZ. This follows the main research question that reads:

How does the economic geography evolve at the rural-urban transition in Bihar, and how does this differ from conventional understandings focused on a central agglomeration?

In order to investigate the developments in the local economic geography, in-depth and short interviews with key informants, small business owners, farmers and students are complemented by field observations in this inductive research design. In particular, I adopt the perception of personal future prospects as a key indicator for construing migration patterns and changes in the composition of livelihoods. The comparison between the findings and prevalent theories is bolstered by an exploration of Muzaffarpur, a Bihari city resembling characteristics of a central agglomeration with rapid growth. The aim of this comparison is to illuminate differences based on the function and size of settlements.

3_{The Bihari GDP per capita in current prices was 36,964 rupees (or about 450 euros) in 2016, roughly one third of the Indian average,}

(10)

10

The upcoming chapter presents the framework of this study, thereby explicating the prevalent and alternative theories for growth in the RUTZ. Chapter 3 introduces the design, data and methods of this study, including an operationalisation of the adopted concepts and the procedure for selecting the cases. The analysis of contextual data, focused on the case area and India’s urbanisation pattern, is found in chapter 4. All collected data in Samastipur is presented and analysed in chapters 5, 6 and 7, followed by chapter 8 that compares this analysis to the contrasting theoretical explanations. Finally, the conclusions, recommendations and limitations are presented in chapter 9. Findings from Muzaffarpur are found in a separate benchmark profile in chapter 4. Profiles with narratives from relevant respondents are presented throughout the chapters in order to illustrate the analysis.

This study’s focus on ill-understood transitions in India’s understudied RUTZ is further motivated by societal relevance. Gaining a better understanding of Bihar’s seemingly different urban system by directly involving local residents is an opportunity to improve living conditions in an area where growth is occurring rapid and often chaotic. I contribute to current and upcoming research that concerns the steering of growth, planning of towns and infrastructure, economic development and current development schemes.

(11)

11

CHAPTER 2

Theoretical Framework

2.1 Introduction

This chapter sets forth prevalent understandings of urban and economic growth in the urban-rural transition zone, which is understood to be an originally rural and poor area now witnessing rapid yet dispersed urban and economic growth1_{. The framework starts with an}

account of the general relationship between economic growth and urbanisation. Then it identifies the key factors that drive growth in the urban-rural transition zone, which are at the same time characterised as either taking an ‘exogenous’ or an ‘endogenous’ perspective, thus distinguishing between external and internal factors2_{. In particular, theories of New}

Economic Geography (NEG) and subaltern urbanisation are introduced for these two different perspectives. A variety of topics such as the role of migration and adjoining capital flows in the urbanisation processes will be delineated. In each section, based on the different theories, a number of hypotheses will be formulated.

2.2 Urbanisation and economic growth: moving in tandem?

This section will expound the general claim that urban and economic growth are moving in tandem. This reciprocal relationship is described by various established theories and empirical studies. Bloom, Canning & Fink (2008) demonstrate the correlation between urbanisation and economic growth and Henderson (2003) suggests an optimal urban concentration grade for enhanced productivity, though hard evidence remains equivocal based on these quantitative analyses. A 2009 IMF report specifically focuses on this relationship and presents cross-sectional data that indicates urbanisation to be a “very strong indicator of all aspects of productivity growth over the long run” (Buckley, Spence, & Clarke Annez, 2009, p. 4).

Traditional theorists of economic geography such as Von Thunen (1966) and Christaller (1933) have selected economic agglomeration and a spatial concentration of people as the key factors for urban growth, and modern adoptions of this can be found in a wide variety of studies3_{. For example, Scott and Storper claim that “cities can be understood in terms of a}

theoretical framework that combines […] the dynamics of agglomeration/polarization, and the unfolding of an associated nexus of locations, land uses and human interactions” (Scott & Storper, 2015, p. 1). Furthermore, Brenner and Schmid explain how traditionally,

1_{A precise conceptualisation is given in section 3.2.2.}

2_{In doing so, this framework follows the working paper by Raman and colleagues on small town dynamics in India that describes}

endogenous factors such as local innovation, as well as the influence of external factors, such as large agglomerations (Raman et al., 2015).

3_{See Brenner & Schmid (2014) for an analysis of studies adopting these factors, and Kloosterman & Lambregts (2001) for a typical}

(12)

12

industrialisation moved in tandem with concentrated urbanisation – although this is, they argue, nowadays complemented by forms of extended and differentiated urbanisation (Brenner & Schmid, 2015, p. 170). The main advantages of an agglomeration are usually ascribed to economies of scale, which can be the reduction of transportation costs, benefits from network effects, and proximity to other businesses and labour pools. (Turok & McGranahan, 2013, p. 467).

The sectoral composition of the GDP seems to change when economies grow – specifically, there appears to be a shift away from agriculture, towards either the manufacturing or the services sector (Buckley et al., 2009, p. 8). Evidence for this from India is found in a study by Himanshu, Lanjouw, Mukhopadhyay & Murgai (2011), who statistically depict how rural poverty reduction and rural non-farm diversification are closely related. Turok and McGranahan enhance these statements by describing the interrelation between the sectoral transition, urbanisation and economic growth: the benefits of agglomeration mainly apply to the manufacturing and services sector, while agriculture is dispersed and intensive in land-use (Turok & McGranahan, 2013, p. 468).

Scott and Storper depict why the clustering of economic activities and labour pools in an urban environment has always been a main reason for urbanisation: “agglomeration occurs because activities like these entail divisions of labor and other interdependencies as expressed in transactional relationships whose costs are distance dependent and because they can reap functional synergies by clustering together in geographic space” (Scott & Storper, 2015, p. 4). The clustering of non-farm economic activities, combined with the statement that economic development is usually not derived from agriculture, is thus a possible causality for the moving in tandem of urbanisation and economic development. Most of the theorists above, however, remain cautious in specifying causal mechanisms: Turok and McGranahan’s meta-study, for instance, finely points out that the "urbanization-growth nexus" is often presented as "some sort of universal law governing an immutable historical process", whereas specific knowledge – especially in Asia4_{– about the dynamics and}

conditions in which these two concepts affect one another remains limited (Turok & McGranahan, 2013, p. 466). And, while stating that the question of causality between jobs and people has no direction, Storper claims that “Deciphering the causes of agglomeration will ultimately require a much better understanding of the complex interdependencies between agents that lead them to congregate together” (2010, p. 322).

2.3 Exogenous drivers of growth theoretically understood

Having established the existence of a general relationship between urbanisation and economic growth, the question now is how different theoretical perspectives describe the driving factors of both urban and economic growth. This section illuminates the exogenous perspective on the drivers of growth, in particular theories of New Economic Geography

(13)

13

(NEG) based in agglomeration and cluster theories that emphasise the benefits of cities for overall economic development5_{(Raman et al., 2015, p. 24). In seeking to understand what}

the location of economic activity determines, a strong interrelation between economic activity and urban growth is assumed. Even in peripheral areas, the main drivers of urbanisation are described as proximity to and connectivity with nearby large metropolitan cities (Fujita, Krugman, & Venables, 1999; Venables, 1994). In conceptualising New Regionalism, Soja (2017) claims that networks of urban agglomerations are defining regions, thereby following Scott and Storper and diminishing rural-urban distinctions. NEG theories thus underline factors external of the growing peripheral town itself and indicate the effects of agglomerations on other areas. The focus on proximity and cluster effects prioritises infrastructure as one of the main variables in interfering the growth-urbanisation dynamic (Turok & McGranahan, 2013). The claim of spillover effects in developing countries from agglomerations to peripheral areas is echoed by a case study of Jodhka & Kumar (2017) in what they describe as "backward" Bihar, although it is unclear whether they follow the NEG argument that the urban growth in the periphery also primarily depends on major agglomerations and the regional economic structure (Raman et al., 2015). While not explicitly associated with NEG, Brenner & Schmid argue on a similar note that emerging urban forms are still to be understood in how they are both dependent on and exploited by larger agglomerations (Brenner & Schmid, 2014).

When it comes to the spatiality of economic activity, most NEG theories generally claim that an open economy6_{leads to a dispersion of economic and urban activity}7_{, although}

several theorists state that cumulative causation of activity leads to unequal dispersion across regions (Fujita et al., 1999; Perroux, 1955). Venables (2005) investigates the spatial path of growth processes by drawing on first and second nature geography, referring to natural advantages and the presence of agglomeration forces that, according to his models, explain inequality among developing countries. He first argues that the positive effect of city size on economic growth (as sustained by, among others, Henderson (2003)) also runs in the opposite direction. While admitting that his models are far from covering all important issues (excluding for instance infrastructural factors), his focus on the spatial and economic effects of agglomeration development on the broader region make it typical for NEG. This also confirms the urbanisation-economic growth correlation.

The prime criticism on NEG has been that its foundation in economics of scale makes it neglect important endogenous drivers of growth, thus overlooking activity in the places themselves when focusing on the rural transition. This criticism has led to alternative theories of subaltern urbanisation (Raman et al., 2015, p. 27). While large agglomerations usually are the subject of most NEG urbanisation research, this is seen as a bias by these other theorists that seek to describe the urban world in its full heterogeneity, and regard the

5_{This school of theory is introduced in section 2.3 and can be traced back to Marshall (1920), a classical theorist on industrial clusters.} 6_{Modern India can be characterised as an open economy (Kohli, 2006).}

7_{An interesting example is Venables's study (2005: 22), who provides several NEG models for - among others - open and developing}

economies. He claims that these provide evidence for the tendency of dispersion of economic activity towards the edges of the economy once the number of manufacturing firms increases.

(14)

14

study of small cities as a crucial element (Bell and Jayne, 2009, in: Mukhopadhyay et al., 2017). Finally, the causal power of NEG theories for explaining the dynamics of agglomerations is contested (Scott & Storper, 2010, in: Turok & McGranahan, 2013). It is thus time to examine these alternative theories of endogenous growth and explore if they offer a better explanation of what is occurring in the rural-urban transition.

2.4 Endogenous drivers of growth theoretically understood

A key difference between endogenous and exogenous growth theories is already given above when discussing the criticism on NEG: while exogenous theories usually derive their arguments from the influence of agglomeration forces, growth can also occur autonomously, depending on local factors, according to endogenous theorists. Moreover, the practice in the RUTZ would require alternative theories that – instead of finding their base in Western geography – are developed in the context of the developments taking place with an endogenous origin.

A variety of theorists and concepts is concerned with urbanisation out of the scope of influence of agglomerations: the urban fringe, peri-urban growth and the general interconnection between urban and rural, often referred to as RUrban (Denis, Mukhopadhyay, & Zérah, 2012, p. 56). For instance, the peri-urban would be “[…] not fixed, rather being in continuous flux and transition” (Ravetz et al., 2013, p. 14) and “blurring the urban-rural distinction” (Ros-Tonen, Pouw, & Bavinck, 2015, p. 86) while the RUrban aims for a “co-evolution of the countryside and of the city” (Revi et al., 2006, in: Denis et al., 2012, p. 57).

The theory of subaltern urbanisation – as proposed by Eric Denis and his colleagues – specifically concerns itself with urbanisation in what I coin the Indian RUTZ. It is their attempt to investigate growth of settlements in cases that are not orchestrated by a specific actor, whether institutional or private, and are beyond explanations of NEG, particularly economies of agglomeration (Denis et al., 2012; Mukhopadhyay et al., 2017). The autonomy of an urbanising entity is a key element in this theory and described as the ability to affect their growth process and interact with other settlements. Further, subaltern urbanisation is different to metropolitan, administrative or corporate urbanisation (Denis et al., 2012, p. 52).

In another criticism of the clear delineation of rural and urban and the focus on agglomeration economies, the Asia-specific theory of Desakota regions as proposed by McGee (1991) offers an alternative. While the peri-urban is described by this theory to be the area where commuters live that work in the urban centres, the Desakota or transformation region stretches further along transportation links (see figure 2.1). Previously, these areas were often densely populated and agricultural, but now know mixed economic activities, partly stimulated by government policies. In an empirical study based on the Desakota theory, rural industrialisation rather than a spillover from agglomerations appeared to push urban and economic growth in China’s Pearl River Delta (Lin, 2001). Useful recent

(15)

15

applications of the theory in India are hardly found, besides studies claiming that the existence of corridor or highway urbanisation is significant (see e.g. Balakrishnan, 2013). Lastly, in his study of reciprocity between villages and cities, Shah (2012, p. 19) illustrates that urban networks exist between small towns and villages without much connection to distant large cities, emphasising again the autonomy of peripheral areas.

Figure 2.1: McGee’s spatial configuration of an Asian region

Source: McGee (1991)

Having established that endogeneity is focused on the autonomy of growth from other actors, I now turn to discussing alternative endogenous drivers of urbanisation and whether people are ‘pulled’ towards settlements or ‘pushed’ away from the countryside. This renounces Raman et al. (2015) who underline the importance of understanding the circularity of migration – rather than the linear migration from rural to urban, which they claim has been the prime occupation of most theorists thus far – for comprehending urbanisation patterns8_.

Ravetz, Fertner, & Sick Nielsen (2013) describe changes in demographics – causing a push towards small towns – together with a growth of production and consumption patterns – usually taking place in towns and pulling people towards them – as most important stimuli of urbanisation in the rural-urban transition9_{. Furthermore, the range of economic and social}

causes of urbanisation as distinguished by Denis et al. indicates migration from the

8_{Similar claims have been made by Shah (2012), Sheppard (2014) and Storper & Scott (2009).}

(16)

16

countryside caused by a pull to health and educational facilities in small towns as most specific and relevant for my study (Denis et al., 2012, p. 58).

An important study of rural livelihoods in Africa, resembling analyses of the Indian reality, shows the significance of diversification as a livelihood strategy. People diversify as they are pushed from agriculture due to population pressure, diminishing returns and “fragmented landholdings” and pulled towards the proximity of urban areas and commercial agriculture (Barrett, Reardon, & Webb, 2001, pp. 315–6). This situation causes some scholars to describe these farmers as the new precariat (Guérin, Venkatasubramanian, & Michiels, 2014).

Nijman (2012) argues that migration towards small settlements is particularly substantial in India's rural-urban migration due to poor agricultural performances, yet simultaneously states that migration is not explaining the entirety of India's urban growth. Mukhopadhyay et al. substantiate both of these claims based on census data, claiming that of the 90.9 million new Indian urbanites in the period 2001-2011, only 19.1 million migrated whereas 40 million urbanites were added due to natural growth and 31.8 million due to "morphing places". This refers to the blurring of administrative boundaries between rural and urban: population growth primarily occurred in Census Towns, small urban settlements without an administrative status, while growth in small towns with an administrative status decelerated10

(Mukhopadhyay et al., 2017, pp. 29–30). Several other case studies find the same trend in other developing countries (see United Nations (2014, p. 24)). A case study of rural-urban migration towards Bangalore produces some insights relevant for understanding the dynamics in the RUTZ, albeit focusing on a megacity incomparable to urban forms in Bihar (Sridhar, Reddy, & Srinath, 2013). It is the hybrid of both push and pull factors that helps construing the reasons behind most migration, and both theoretical models as well as empirical findings indicate urban-rural differences in expected income – rather than actual income differences – as most important explanation for migration. This is in line with Lambin et al. (2001), who depict from several cases that people’s responses to economic opportunities are the main driver of urbanisation.

The theorists above have mainly provided general, structural explanations and refrained from a detailed analysis of endogenous economic development, acknowledging the existence of knowledge gaps here. In order to enhance our understanding, I will now turn to an analysis of the type of economic activities taking place in the RUTZ.

After describing their earlier attempt of providing a classification of urbanising settlements11_{as leading to oversimplifications, Mukhopadhyay et al. now identify two general}

types of settlements: (1) "Micro Global Settlements" that are nationally and globally connected and (2) locally connected towns with many non-farm workers (2017, pp. 31, 35– 6). The first type – a site of production with qualities comparable to large agglomerations –

10_{Section 4.1 expounds on the administrative and census status of a city in India.}

11_{Spatial proximity and administrative recognition are the two axes along which subaltern urbanisation initially classifies settlements}

(17)

17

appears to resemble characteristics of exogenous growth12_{and is not relevant for my study as}

I attempt to clarify growth in the least expected conditions of the RUTZ. The second type is described as a space of change with people drifting between farm and non-farm work; this could be service towns, markets, local transit hubs, and private health and education centres. Its existence in India is made possible by the country’s high population density and improving rural infrastructure. Three sorts of economic transformation for this settlement type are identified: (1) the everyday or bazaar economy that serves the surrounding areas, does not generate surplus capital, benefits from increased consumer's demand and adds employment through logistics; (2) small-scale settlement-specific activities, sometimes linked to natural resources, depending on improved infrastructure to obtain increased production; (3) a wide variety of new activities, mainly construction and para-transit services, but also services related to electronics (e.g. cell phones), private education and private healthcare. Public investment, increased demand for travel and, crucially, providers of finance for the purchase of vehicles are key to increase particularly transit services (ibid., pp. 33-4).

More evidence about the economy in this type of locally connected towns is provided by Jodhka and Kumar’s study (2017) on changing economies in Bihar, mapping the non-farm economic activities in two settlements. They conclude that none of these activities are characterised as modern manufacturing but most are rural, non-farm activities that cater for local demand, which is in line with economic data and resembles the bazaar economy. The significance of new activities – as Mukhopadhyay et al. describe – for the RUTZ’s local economy is not given in this study, however. The only substantial source of modern and formal employment is found in institutions that provide services such as schools, banks and government offices; this is an external source of growth (Jodhka & Kumar, 2017, pp. 19, 21).

At the same time, government data reveals that Bihar’s countryside has witnessed a sharp increase in non-farm economic enterprises13_{, suggesting a significant and expanding}

source of informal employment for the RUTZ. Jodhka and Kumar’s case study finds livelihoods under pressure; in particular farming is not seen as a “viable option” even by the few farmers who own large land plots (ibid., 19). The above provides evidence for the statement that economic diversification and sectoral change towards non-agricultural economic activity is the only strategy for finding a livelihood, and the authors conclude that the two cases "[do not] support the popular assumptions about the urbanising effects of a growing non-farm economy" (ibid., 22) thereby hinting that these would be economic, market-driven growth.

Lastly, a number of theorists identify remittances – capital sent from a migrant worker in a large metropolis to others, usually family members, in the hometown – as one of the drivers of growth, although little is known of the phenomenon in RUTZ (Denis et al., 2012). A study about the sources of growth in the Indian state of Kerala found that remittances are

12_{It can be seen in line with how Ros-Tonen, Pouw, & Bavinck (2015) characterise urban-rural linkages: as metabolic flows of people,}

materials and more, and influenced by global connections, thereby following theorists such as Castells (1972).

13_{The number has increased by 40% between 2005 and 2013, adding 500,000 new enterprises to Bihar – see chapter 4 for more}

(18)

18

a key factor in the increase of demand in services and "durable goods", stimulating economic growth in the region (Pushpangadan, 2003). In 2011, 78 percent of all households in a typical Bihari village had a migrant worker, usually male and in half of the cases below 30 years old (Datta, Rodgers, Rodgers, & Singh, 2014). Interestingly, much of this migration would be circular and migrants would maintain a strong connection with their hometown. Further stressing these connections, small entrepreneurs would obtain their funding from remittance flows (Raman et al., 2015, pp. 38, 46), Datta (2016) even claims that remittances are the lifeline in Bihar, with 29 percent of all households relying on it and a study of rural-urban migration in Karnataka emphasises that it is often a family decision (Sridhar et al., 2013, p. 289). One could argue that these remittance flows are another exogenous driver of growth, although the apparent strong connection between migrants and their hometowns with relatives dilutes this argument. It is instead characterised as a general driver.

2.5 Conclusion

The theoretical framework provided numerous interpretations and explanations for the drivers and dynamics of growth in the rural-urban transition. This generates eight literature statements from a general, endogenous or exogenous perspective:

• General Urbanisation and economic growth generally progress in tandem (IMF, 2009; Bloom et al., 2008; Brenner & Schmid, 2015; Henderson, 2003; Von Thunen, 1966).

• General Economic growth in rural Bihar will not have agriculture as its primary cause but is rather based in other sectors (Buckley et al., 2009; Scott & Storper, 2015; Turok & McGranahan, 2013).

• General

The financial and emotional connection between a migrant and its hometown in rural Bihar remain strong, supporting the economic development of these settlements (Datta, 2016; Denis et al., 2012; Raman et al., 2015).

• Exogenous Even in Bihar’s rural transition, urban and economic growth is still related to nearby agglomerations (Brenner & Schmid, 2014; Fujita et al., 1999; Marshall, 1920; Soja, 2017; Storper, 2010).

• Exogenous

Agglomeration forces and natural advantages cause economic activity to be unequally distributed across regions and predominantly focused in cities (Fujita et al., 1999; Henderson, 2003; Perroux, 1955; Venables, 2005).

• Endogenous Growth in the urban-rural transition is occurring largely autonomous from important nearby agglomerations (Denis et al., 2012; Mukhopadhyay et al., 2017; Shah, 2012).

• Endogenous Urban growth in the RUTZ can be partially explained by natural growth, and partially by a number of driving factors that all relate to being pushed away from agriculture and being pulled towards amenities in a

(19)

19 settlements (Mukhopadhyay et al., 2017; Nijman, 2012; Raman et al., 2015; Ravetz et al., 2013).

• Endogenous Economic growth in Bihar’s RUTZ originates mostly from local, informal, non-farm, consumption-focused economic activities (Jodhka & Kumar, 2017; Mukhopadhyay et al., 2017).

My conceptual model departs from the assumption that urban and economic growth are in some way related to each other, but that the strength, direction and conditions of this relationship remain contested or unknown and are context-dependent. This especially applies to India’s RUTZ, a highly rural and economically undeveloped yet rapidly growing area. Theories of New Economic Geography (NEG) and subaltern urbanisation provide contrasting interpretations for the drivers of growth here. Exogenous NEG theories based in agglomeration economies appear to have been the dominant view in urban growth studies, though their validity decays when moving further away towards remote areas in the Global South that are ill-connected with larger cities, giving way to endogenous, subaltern theories and inductive research. In particular, my conceptual model identifies the autonomy of a settlement as defining the validity of exogenous or endogenous theories.

Despite the significance of agriculture for the area, it seems unrealistic – both from an endogenous and an exogenous perspective – that urban and economic growth in the rural-urban transition originate from agriculture. Instead, an alternative driver of economic growth as identified by scholars is found in self-providing services started by individuals, possibly funded in part by remittances, though strong evidence for this is still scarce. People could furthermore be pushed out of agriculture and small villages, yet unable to move to even larger agglomerations, and pulled by the amenities conditions of living in the nearby smaller towns. These key concepts direct my model towards non-agricultural economic drivers of urbanisation.

The field of study still faces many gaps according to several theorists14_{. Among others,}

these include the influence of exogenous factors on the local development, the urban network of small towns, the livelihood strategies of residents, the factors that influence the location choice of firms and the spatial practice of local economies. The above motivates the aim of this study to offer a highly explorative account of the recent urban and economic dynamics in the Bihari urban-rural transition, with a focus on the concepts identified above.

(20)

20

CHAPTER 3

Research Design

3.1 Introduction

Now that I have established the aim of this study within its theoretical context, I turn to the design of the study. This chapter provides the overall design and the research questions, operationalisation, case selection, data collection, methodology and positionality.

This study is developed under the scope of the NWO-funded research project “A New Urban Epoch? Alternative ways to measure urbanization using VHR remote sensing methodologies” by Robbin Jan van Duijne at the UvA. I executed my apprenticeship – as part of the Urban Studies master programme – within this project by working on new measurements of urbanisation. Especially due to the limited research available, Van Duijne and I sometimes cooperated while developing research ideas and conducting contextual analyses. In particular, resemblances can be found regarding the overall aim of our research. Some content in this thesis has been derived from Van Duijne’s research project, and this will be referred to in the text. However, all research and decisions are mine and I conducted the fieldwork alone1_.

3.2 Research questions and operationalisation

3.2.1 Research questions

This case study of new expressions of urbanisation focuses on the dynamics of economic geography in the RUTZ, an area that rapidly develops yet faces significant knowledge gaps2_.

The theoretical framework has delineated relevant hypotheses based on both traditional as well as alternative theories. The suitability and relevance of conventional theories – usually focused on agglomerations for explaining the concomitant growth of economies and cities – for the developments in the RUTZ is contested. It is my aim, therefore, to firstly provide inductive and exploratory research on the ill-understood dynamics, and to secondly assess whether conventional or alternative theories hold value for explaining the developments in the case area. In line with this aim, the main research question reads:

How does the economic geography evolve at the rural-urban transition in Bihar, and how does this differ from conventional understandings focused on a central agglomeration?

1_{Robbin Jan van Duijne selected the same settlement as my main case as one of the research sites for his project, and visited the area}

several months after my fieldwork period.

(21)

21

This main research question is divided into three sub questions:

1. How does the agricultural economy evolve at Bihar’s rural-urban transition? 2. How does the urban economy evolve at Bihar’s rural-urban transition?

3. How do residents perceive their future prospects at Bihar’s rural-urban transition?

My investigation of the agricultural and urban economy is focused on changes in economic activities, employment structures, livelihoods and the regional function, which is together understood as the economic geography. The first two questions combined will illuminate developments and transitions in the entire local economy as well as clarify how the two sectors relate to each other. This analysis subsequently aims to construe the economic push and pull drivers of urbanisation. The third research question then turns to the residents’ perception of the future which will firstly lead to the inclusion of the residents’ perspective on the developments they witnessed directly into this study. Secondly, investigating the residents’ perception on different personal future scenarios – when staying in Samastipur and when migrating away – will clarify the impact of migration on the growth of the city, and the push and pull drivers that cause this growth.

3.2.2 Operationalisation

Studying drivers of urbanisation firstly brings forth the question of what constitutes the urban3_{. It is beyond the interest of this study to examine the difference between urban and}

rural, therefore no specific operationalisation of the city is provided. Rather, it is seen as the outcome of the process of urbanisation4_{. Nonetheless, existing understandings of these}

concepts as well as their impact on different actors (e.g. migrants, policymakers, local businesses) cannot be neglected. I thereby choose to follow Mukhopadhyay et al. (2017: 30) who claim that binaries between urban and rural have effect on reality and influence the actions of the actors involved. Furthermore, I adopt the definition of settlement to describe the adjacency of built-up area, regarded by people as a village, town, or city. An agglomeration is a settlement bearing agglomerative functions (as described in the theoretical framework).

The locus of this study, the rural-urban transition zone (RUTZ), follows the notions of subaltern urbanisation (cf. Raman et al. (2015, p. 27), among others) and locally connected towns (cf. Mukhopadhyay et al. (2017)). The zone’s location is out of the direct scope of influence of a larger agglomeration and faces economic and socio-cultural transformations. Socio-cultural characteristics of an urban-rural transition are less apparent, more difficult to define and not the focus of this research. Therefore, no characteristics will be pre-defined, although the respondents’ perception on socio-cultural change may be included in this study.

3_{For an account of this study’s stance regarding epistemology and ontology, see section 3.3.2.} 4_{This follows the arguments of sociologists such as Harvey (1996) and Wirth (1938).}

(22)

22

Changes over time in the administrative definition of a settlement from ‘rural’ to ‘urban’ are a clear indicator of rural-urban transition. In India, two urban definitions can be distinguished: the administrative definition of ‘statutory towns’ (STs), issued by state governments and determining the financial resources of the urban body, and the census urban definition. The latter comprises three criteria: a population of over 5,000 inhabitants, a population density of at least 400 persons per square kilometre, and at least 75% of the male main working population5_{engaged in non-agricultural economic activities. Although both}

definitions are very stringent as compared to other countries, thus possibly misinterpreting the real rate of urban growth6_{, the inclusion of the socioeconomic criterion in the census}

definition allows for examining both the economic and geographic transition and is thus adopted in this study for that purpose.

The RUTZ also faces a transformation in the local economy, which is defined as the system of economic activities within the settlement. The first two sub questions identify two different economic sectors: agriculture and the urban economy, requiring a distinctive operationalisation. The first is primarily focused on agriculture, while the RUTZ witnesses a transformation towards more developed and mixed economic activities, resembling McGee's notion (1991) of a transformation or Desakota zone. The share of employment in agriculture in the local economy should thus be declining over the years. Agriculture can be seen as all economic activity producing or processing agricultural goods, in line with the census definition of ‘agriculture and allied sectors’, in particular comprising people working as ‘cultivator’ and ‘agricultural labourer’. Additionally, I perceive activities directly related to agriculture – such as basic processing of originally agricultural products for other purposes, e.g. handicraft – as agricultural as well.

The urban economy comprises all economic activities in the RUTZ which are not directly related to the production or processing of agricultural goods. According to census definition, these comprise ‘household industry worker’ and ‘other workers’. Changes in the number of people employed in these sectors is a direct indicator of the shift from rural to agricultural and can be derived from census data. Further, observing and analysing economic activities is based on the framework of everyday and settlement-specific activities of Mukhopadhyay et al. (2017) with a focus on consumption-focused, self-employed businesses. This analysis includes the land use and built environment; geographical components which are used for characterising activities in and outside the settlement. Further indicators for the transition and drivers of the local economy are provided by the level and composition of livelihoods in different sectors, including possible remittance flows, along with overall levels of purchasing power and consumption. Data for this is collected directly from respondents at different locations.

5_{‘main workers’ is the accepted definition for measuring the workforce using the Indian census. It excludes marginal workers, people}

who worked for less than half of the year.

6_{Different scholars have developed alternative definitions and/or calculated India’s urbanisation rate based on the same census data.}

Most of these definitions provide urbanisation rates varying around 50%. See, for instance, Denis & Marius-Gnanou (2011); Jana, Sami, & Seddon (2014); Tandel et al. (2016).

(23)

23

Turning to the drivers of economic growth, I make a distinction between endogenous or exogenous drivers. Key to distinguish between these is the autonomy of the settlement, which is defined as "[a settlement’s] ability to affect its growth process and interact with other settlements" and being largely independent from a major agglomeration (Mukhopadhyay et al., 2017, p. 28). The identification of the main pillars of the local economy by various respondents will be analysed and some indicators are relatively evident: a large share of residents being employed by an outside employer, or spillover effects from nearby larger cities, for instance. The regional function of the settlement and its economic activities can clarify ‘pull’ effect of the local economy on both economic and urban growth. Concerning the drivers of urban growth of the settlement, this study focuses on the influence of developments in the local economy and on migration flows7_{. Investigating}

respondents’ places of origin and current residency is done to illuminate existing migration patterns. In order to enhance the analysis of the push and pull drivers behind migration, the perception of residents and students on their prospect regarding different future scenarios of employment and residency is investigated. This future prospect is also adopted for enhancing the understanding of livelihoods, assuming that the perception of the future situation is influenced by the current situation.

3.3 Comparison and cases

This case study focuses on the unexplored causes of growth in Bihar’s RUTZ, exemplary for a deviant practice of urbanisation which is challenging current theories of urbanisation8_(Yin,

2013, p. 39). Following a contrast-oriented logic, the aim is to generate new hypotheses (Gerring, 2006). The unit of analysis is the rural-urban transition zone.

After concentrating on in-depth research in the main case, inductive in nature though embedded in a framework of subaltern theories, the study makes a comparison with a framework of currently prevalent models and theories of urbanisation (or, as Gerring describes, an 'informal comparison' (2006, p. 106)). The prime focus of this comparison is on the assumptions of the theories and the causality and conditionality of the urbanisation processes, instead of the accuracy of their predictions (following the systematic process analysis posed by Hall (2003, pp. 21–3)). Furthermore, a concise investigation of a case representing benchmark or ideal type of these agglomeration focused models within Bihar is included in order to bolster this comparison (Skocpol & Somers, 1980, p. 178). Besides enhancing the validity of the main case research, this should help illuminate whether the main case actually represents a deviant example compared to the traditional framework, and if so, to what extent and on what aspects. The data collection in this second benchmark case is limited. Triangulation of methods – mainly through combining different types of interviews with observations – suits this exploratory design and stimulates reliability and

7_{As illuminated in the theoretical framework, natural growth additionally forms a significant factor for the growth of settlements. Migration}

is, however, regarded as fluctuating more and thus has more potential impact on the growth of the city.

(24)

24

generalisability of the results (Yin, 2013, p. 29). Finally, the inductive nature and the direct inclusion of the residents' perspective in a rapidly developing though still ill-understood area allows this study to be transformative (Burawoy, 1998).

The main case is representative of a pattern of urbanisation that is assumed to deviate from traditional patterns but instead comprises characteristics of ‘subaltern’ urbanisation. The Indian state of Bihar is selected for this for a number of reasons. There are indications that India’s urbanisation is increasingly occurring in smaller settlements instead of megacities (see section 4.1). The poor and extremely rural yet densely populated state of Bihar, consistently ranking as one of India’s lowest on various development parameters, is witnessing rapid economic growth (see section 4.2). Bihar thus questions prevalent theories that link development to the growth of large agglomerations, and is often neglected in the research bias towards large cities. Besides this theoretical relevance, selecting Bihar also enhances this study’s societal relevance: my policy recommendations may contribute to improving the precarious situation of development in the state.

In order to assess the extent to which agglomeration economies still influence growth in the RUTZ, a typical RUTZ case is selected. This Bihari settlement is located as far as possible out of the direct influence of such an economy and is not well connected in general. At the same time, the settlement is facing growth in both economic and population terms at high rates, around the Bihari average, especially regarding economic enterprises that can be seen as urban economic activities. Feasibility was defining the case selection as well. I had never visited the area before, needed to be able to find a local assistant, and some parts of the state face political unrests. I thus decided the settlement’s population should amount between 50,000 and 100,000, also making sure that the settlement is not resembling a large agglomeration.

The conditions above have been applied to the latest available census data (2001 and 2011) and I undertook a geographical analysis of urbanisation in Bihar. Consequently, the city of Samastipur was selected for the main case. This City Council9_{had a population of}

67,925 in 2011, although the adjacency of a significant number of growing census towns – with a large share of the population active in the non-farm economy – implies the existence of a dispersed, larger settlement. Population and economic growth are around the Bihari average, depending on the adopted definitions. Road connectivity to larger cities is limited, presumably taking almost two hours to the larger city of Muzaffarpur and twice that time to the state capital Patna. Samastipur is better connected by rail: Muzaffarpur is reached in 90 minutes with a commuter train (interview 1) and direct connections to other cities exist and the rest of the country exist too, although a direct line to Patna does not exist10_.

For the benchmark profile of a typical case of agglomeration-focused growth, Muzaffarpur is selected. With a population of 354,462 in 2011 for the Municipal Corporation9_{, it is much larger than Samastipur, and the presence of the national highway as}

9_{These are administrative urban definitions in India, taking into account the size and function of an agglomeration.} 10_{A detailed analysis of Samastipur’s urbanisation and connectivity is found in section 4.4.}

(25)

25

well as a direct train connection to state capital Patna make it better connected. Muzaffarpur is surrounded by a concentration of census towns and nonfarm villages which appears to be the highest in Bihar (see section 4.3). The relative proximity to Samastipur allows for conducting research with the same assistant, and some contextual factors – climate and culture, for instance – are comparable as well.

3.4 Data collection, analysis and methods

3.4.1 Collection

Given the little amount of research undertaken in Bihar and the lack of relevant and detailed available statistical data, this study is highly dependent on the exploratory collection of data. Following a multi-method approach, the focus of data collection is on structured, semi-structured and in-depth interviews and observations. All data has been collected during a one-month period of intensive fieldwork in February 2018. Albeit preparations at IIT Bombay11_{for this period, only on the site I managed to recruit local assistance using UvA}

funding. Bhushan Kumar was working in several Bihari NGOs and as an assistant he supported this study during explorations of the area and through finding contacts in the first week. Mrityuntjay Jha, a local English teacher, served as translator in Hindi and Maithili, and thereby attended and supported most of the data collection.

Structured and semi-structured interviews have been conducted with 72 residents for the main case, categorised in three general groups (see table 3.1 for basic characteristics and Appendix 3 for an overview of all respondents). Farmers and business owners are selected in order to illuminate economic activities, livelihoods, employment, capital, and future prospects. The inclusion of students into this study primarily served to enhance the analysis of future prospects and additionally aimed to construe the role of educational facilities as a potential pull factor for migration. Selective sampling is adopted to obtain a diverse sample representative for 1) both old and new neighbourhoods 2) different types and sizes of economic activities, with a focus on self-employed businesses and 3) almost finished students at different levels, in both genders.

After piloting different interview forms, semi-structured interviews fit best in my inductive design. Respondents are less biased and influenced in a least-structured environment. The interviews with business owners and farmers are structured around a list of questions for most indicators complemented by some specific background questions12_and

all answers are summarised by me on paper or on my phone. Because of the narrow scope of questions for students combined with their higher willingness to cooperate, these interviews were shorter and conducted in a very structured manner. All interviews were conducted by me and accompanied by the assistant, who translated all questions and answers precisely.

11_{This was the university I attended for my exchange as part of the Urban Studies programme. On this institute, I spoke with various}

students and lecturers to prepare my fieldwork; Appendix 4 contains a list of my contacts at IIT Bombay.

(26)

26

Table 3.1: Participants of semi-structured and structured interviews

Group (n) and method Location (n)

Students (24): structured

Prerna Computer School, (7)

Genius Commerce School, Kashipur (10) Private school M. Jha, Kashipur (7)

Business owners (37) of which 11 large businesses: semi-structured

Central (10)

New areas Kashipur (10) Bazar Samiti (12) Other (5)

Farmers (9): semi-structured Bangali Tola market (9)

Further data collection consisted of 13 in-depth interviews with key informants, structured around a topic list (see Appendix 2). Some of the respondents were pre-defined: I conducted interviews with a number of different officials in order to include the administrative perspective on the economic-geographical dynamics (see table 3.2, primarily respondents 1 to 5). Additionally, in-depth interviews with key informants are meant to illuminate more about the developments over the years. These respondents are people with substantial local knowledge and experience about the topics concerned. They were selected while being in the field, and the respondent was usually acquainted by originally taking part in the short interviews or being familiar to one of these respondents (see table 3.2, primarily respondents 6 to 13). In selecting the 13 respondents, I attempted to include as much a variety of perspectives as possible. Almost all interviews were attended by the translator, even though the interviews were mostly held in English, and 6 interviews are recorded in order to enhance the processing and analysis of the collected data. For the other interviews, recording was not possible. All participants gave their consent for participating in my study and, if applicable, for recording the interview.

The last part of data collection consists of observations and photos made by me. This concerned the road and rail infrastructure, built-up typology and connectivity of the agglomeration, which are mainly for the context of the analysis, but also the economic activities. A number of locations with a concentration of businesses in both the old and new, urbanising parts of town was selected for analysing the different types of economic activity in Samastipur. During various visits to the surrounding areas, I observed and photographed the economic activity visible there. Photos and observations regarding connectivity were made at different times, during and after peak hours, and repeated at key locations in the infrastructure of town. Additionally, I undertook a focused group discussion with students at Genius Commerce School, structured around the same topic list as for the interviews, and observed a meeting of starting business owners. The main source of secondary data is the Indian census, conducted every ten years. Although the census collects data on a wide range of variables, only a limited selection is available for the district and sub-district level. In particular, I used the 2001 and 2011 round with a focus on population growth.