Economic Liberalisation,Transport Cost and In-situ Urbanisation in India: A General Equilibrium Analysis

Abstract

Urbanisation in India is moving away from metropolitan dominance and showing a different trend recently. It takes a new path due to the growth of numerous small and medium urban bodies through reclassifying villages, popularly known as Census Towns. Using a general equilibrium framework, the paper finds that, due to economic liberalisation and an increase in capital inflows, the process of reclassification-based urbanisation tends to accelerate. However, the net effect of this type of in-situ urbanisation is ambiguous due to increased transportation costs. The empirical observation confirms the theoretical findings. We find a positive relationship between urbanisation and the inflow of foreign capital, proxied by the number of banks in the nearby city. A negative relation is observed when we proxy the transportation cost by the distance from the nearby highway, and a positive relation is found when the same is proxied by the distance from the nearby city.

JEL: D58, J01, J60, R10

Keywords

Census town capital inflow transport cost rural–urban migration

I. Introduction

The Indian story of urbanisation does not conform to the global trend. According to official statistics, India’s migration has never reached the expected pace. It has remained almost stagnant, accounting for one-fifth of total urbanisation for the last three decades. The pattern of urbanisation in India shows a different trend, with new urban growth coming from metropolitan dominance. Out-migration from big cities has become a significant feature in recent times.

Kundu (2011a) demonstrates that the metropolis and large cities like Delhi, Chandigarh, Kolkata, Hyderabad, Chennai and Mumbai have experienced a decline in the growth of population in the last decade, owing to the decrease in in-migration to large cities due to formalisation, which makes these urban centres less attractive to the marginal section. The formalisation and sanitisation, like removing encroachments, slums, petty commercial establishments, squatter settlements in big cities and judicial intervention to remove undesirable urban growth, make poor people more vulnerable to staying there. Bhide (2013) points out that poor migrants, who migrate to urban centres for petty jobs, are becoming increasingly vulnerable due to being denied access to essential services. However, low investment in agriculture and rural infrastructure and premature economic liberalisation often have a mixed effect on migration dynamics.

Figure 1 shows different components of urban growth in India for the last five decades. The natural increase in population always remains the primary driving force in urbanisation in India. However, migration-induced urbanisation has remained almost static over the past five decades, though the country has undergone significant economic changes, such as economic liberalisation and infrastructural development. Another interesting feature is that the reclassification-driven urbanisation has outpaced migration-led urbanisation in the last Census of 2001–2011, owing to the growth of numerous Census Towns (CTs).

A CT is a village reclassified as a town by the Census if it satisfies the following three conditions: (a) population of 5,000 or more, (b) population density of 400/km² and (c) at least 75% of the male main workforce works in the rural non-farm sector. However, a CT is administered by the rural administration (panchayat). Figure 2 shows the distribution of CTs in several decades compared to Statutory Towns (STs), which are the formal urban centres administered by municipalities, municipal corporations and cantonment boards.

As evident from Figure 2, the number of CTs in India has increased from 1,362 to 3,894 during the decade 2001–2011, accounting for almost 30% of urbanisation, as evident from Census 2011, though in the previous decades, CTs did show a marginal growth trend, including the decade preceding 2001–2011.

The literature related to CTs revolves around: (a) the unprecedented increase in CTs during 2001–2011, (b) the spatial distribution of CTs and (c) the role of transport infrastructure and village-specific non-tradables in CT dynamics. As expected, the growth of CTs has lured the attention of economists, geographers, other social scientists and policymakers for possible explanations.

The literature on the unprecedented increase in the number of CTs devotes much attention to whether there is over-activism by the Census to exaggerate the data on CTs. Kundu (2011b) initiated a debate by stating that the phenomenal increase in CTs is an outcome of activism by the Census. Because of the enthusiasm of state directorates in identifying the urban centres, the Registrar General of India (RGI) was under tremendous pressure to review the methodology of data collection for identifying the urban settlements. However, Guin and Das (2015) statistically show no such Census activism, as Kundu (2011b) opined in the West Bengal context. Most of the villages identified as ‘would-be CTs’ in 2001 were converted to CTs in 2011, and all over the state qualified as CTs in 2011. Since agriculture has become non-remunerative in the post-reform period, sectoral diversification is necessary for small and marginal farmers to come out of agriculture and join non-farm activities. Again, the growth of the informal economy and small-scale industry in villages is also responsible for the change in employment patterns, which leads to the development of CTs.

The literature on the spatial distribution of CTs finds that the CTs spatially grow both around existing urban centres and away from them. Pradhan (2013) has shown that the spatial distribution of CTs varies significantly among the states. Some states experience the rapid growth of CTs around existing cities (e.g., West Bengal), whereas in some other states, it grows far away from existing urban centres (e.g., Kerala). According to Pradhan (2013), there is a positive association between the degree of urbanisation of a district and the magnitude of the transformation of villages into CTs in those districts.

How transport costs affect urbanisation dynamics is depicted in the pioneering paper by Krugman (1991) on economic geography, arguing that lowering transport costs accelerates the centripetal forces that attract labour and capital to the city. However, Helpman (1998) points out that the rising cost of non-tradable commodities in the town, like the price of housing, a clean environment and congestion-free traffic, can disperse the centripetal forces mentioned by Krugman (1991). While in the former, a lower transport cost due to improved infrastructure promotes the growth of an existing agglomeration with higher welfare of its residents at the cost of the smaller agglomerations in its neighbourhood, in the latter, a smaller agglomeration grows in terms of population with a falling welfare level of its existing residents. In the context of the US, Chandra and Thompson (2000) show that the highways built between 1969 and 1993 raised the income of the rural counties through which they pass at the cost of a reduction in the income of the adjacent counties. In a recent chapter, Baum-Snow et al. (2020) show with Chinese data that investing in local transport infrastructure to promote the growth of the hinterland often has a self-defeating impact of losing economic activities and specialisation in agriculture in those areas.

In the Indian context, Sharma (2013) points out that improved transport infrastructure plays a vital role in the birth of the CTs near existing urban centres, as workers prefer to commute to jobs rather than migrating to the cities. Ghani et al. (2012) found that district-level infrastructure partly facilitates the relocation of organised manufacturing to rural locations while unorganised manufacturing migrates to urban locations. Such movement is partially explained by the development of national-level highways, especially the construction of the Golden Quadrilateral. Aggarwal (2018) takes a village-centric view and shows that expansion of rural roads and transport infrastructure results in greater market integration, reduced price of non-local goods and a wider variety of consumption baskets in rural areas. It also leads to greater participation of local teenagers in the expanded labour market. Mukhopadhyay et al. (2016), with an investigation in certain CTs in northern India, show that increased connectivity and growing rural income are the main driving forces for the growth of small-scale non-tradable services, which are the main sources of non-farm employment in these settlements. Van Duijne and Nijman (2019) note that in the states of West Bengal and Bihar, there has been a shift in the employment structure away from agriculture, causing in situ urbanisation in a dispersed manner. They identify the absence of sufficient employment opportunities in existing urban settlements or major cities as one of the reasons behind this dispersed pattern of urbanisation. However, none of these papers discusses the role of transport infrastructure in the birth of the new CTs as we do in the present chapter.

Balakrishnan (2013) with two examples of the Bangalore–Mysore highway and Pune–Nasik highway, demonstrates that urbanisation along highways has been the emerging pattern of urbanisation in developing countries. Based on NSSO 55th round (1999–2000), 66th round (2009–2010) and 68th round (2011–2012) data, Mahajan and Nagraj (2017) argue that the road construction projects undertaken in India during 2000–2012, both highways and rural roads, expanded rural construction demand and employment.

The impact of economic liberalisation on the reclassification based on in-situ urbanisation has not been studied thoroughly in the context of India. However, economic liberalisation and its impact on the labour market and economy have been studied extensively in the context of the third world, including India. Marjit and Kar (2015) show that reform in the labour market in an open economy leads to the expansion of the urban formal sector and contraction of the urban informal sector, with a fall in wages in the urban informal sector and a rise in wages in the rural informal sector. This causes a reverse migration in the economy with a fall in the average wage until there is substantial growth in the urban organised industry. Krugman and Elizondo (1996), with the example of Mexico, draw attention to an interesting feature that gigantic cities like Mexico City in developing countries are an outcome of import substitution policy when the manufacturing output serves a small domestic market as a consequence of strong forward and backward linkages and consequent agglomeration economies. However, as the economy moves out from an inward-looking strategy through trade liberalisation, the manufacturing sector disperses with the mitigation of backward and forward linkages in Mexico City to northern states closer to the US border, causing the dispersion of the non-farm labour force in the economy. Zhu (2017) identifies the phenomenon of in situ urbanisation in a wide range of areas in southeastern coastal provinces of China, demonstrating that such a phenomenon has been one of the major characteristics of China’s urbanisation process after the 1970s, following liberalisation in line with Krugman and Elizondo (1996).

India, as a capital-scarce economy, adopted liberalised investment strategies to solve certain problems like economic growth and unemployment. Again, a notable feature of India is the growing significance of the rural non-farm sector, which provides an alternative livelihood to the rural masses and attracts the attention of researchers and policymakers. However, economic liberalisation, the inflow of capital, its role in creating rural non-farm activities and the consequent reclassification of villages into CTs are very novel in the literature. Rising urbanisation, liberalisation and improved transportation networks integrate rural and urban economies through economic linkages, providing alternative livelihoods. In this paper, we intend to examine whether the birth of CTs is an outcome of the growth of the rural non-farm sector due to economic liberalisation and consequent capital inflow. Also, we want to study how the change in transport cost with the growth of transport infrastructure has an impact on rural non-farm employment and consequent growth of CTs in a four-sector Harris–Todaro-type general equilibrium framework with three rural sectors, namely the rural farm sector, rural non-farm sector type 1 and rural non-farm sector type 2 and the urban formal sector. We have studied the impact of capital inflow and improved transportation on the rural non-farm sector and the consequent growth of CTs. Though the inflow of foreign capital tends to accelerate CT dynamics, we have found that improved transportation has mixed effects.

A vast literature describes urban unemployment using the Harris–Todaro migration equilibrium condition within a competitive general equilibrium framework. Gupta (1993) first formalised the simultaneous existence of the informal sector and open unemployment in urban areas. Later, Chaudhuri (2000) and Chaudhuri et al. (2006) have also explained open unemployment in the urban industry despite the presence of the urban informal sector. Chaudhuri and Mukhopadhyay (2009) analysed in detail the introduction of the urban informal sector and unemployment. Khan (1982) and Chandra and Khan (1993) attempt to examine the welfare impact of foreign capital inflow using a Harris–Todaro (1970) framework. Chaudhuri et al. (2024) describe rural–urban migration with heterogeneous firms and heterogeneous labourers. However, none of the previous papers analyses the impact of foreign capital, capital inflow and transportation cost on city dynamics and urban unemployment.

This paper has four sections. Section I introduces the topic and explains the phenomenon with the help of existing literature, Section II sets a theoretical model and tries to prove all the propositions mathematically and logically, Section III provides empirical testing of the model and Section IV concludes the paper.

II. Theoretical Framework

The Description

The model developed in this paper is a four-sector small open economy with two internationally traded goods sectors and two internationally non-traded sectors. We consider four factors of production: labour, domestic capital, foreign capital¹ and land. Sector 1 produces a traded agricultural good using land and labour. Sector 2 produces non-traded, non-farm goods. Sector 3 is also a non-traded non-farm sector. Sector 4 produces a traded manufacturing good. Sectors 1, 2 and 3 are located in rural areas, whereas Sector 4 is located in an urban area. Sectors 2 and 3 are both rural non-farm sectors, but they are differentiated in terms of input usage. Sector 2 may include rural sectors like power loom, rural construction, animal feed, fodder industries and brick industry, which use land, labour and capital of a domestic type. On the other hand, Sector 3 may be rural transport, which mainly uses labour as an input. Rural labour is perfectly mobile between Sectors 1, 2 and 3. The rural–urban migration mechanism is of the Harris–Todaro (1970) type. The urban manufacturing sector, which uses labour and foreign capital, faces a protected labour market with a fixed wage rate, but the common wage rate in the three rural sectors is flexible and market-determined. In the migration equilibrium, the expected urban wage equals the rural wage. Land can be used between the agricultural and non-farm sectors of type 1 (Sector 2). Domestic capital is specific to the non-farm sector of type 1 (Sector 2). Sectors 1 and 2 have common inputs of labour and land, and we assume that Sector 2 is labour-intensive compared to Sector 1. Foreign capital is specific to the manufacturing sector (Sector 4).² The price of the traded goods is exogenously given, and the seller’s effective price of this good is changed due to an exogenous change in the policy parameter. The equilibrium price of the non-traded sectors is determined by the equality of their supply and demand in the competitive home market. Demand for non-farm goods depends on their price. The price-demand relationship is negative in each market. The production function of each of these four sectors satisfies all standard neo-classical properties, including constant returns to scale. Factor endowments are exogenously given in the static model. Factor prices, except labour, in each of these three sectors are perfectly flexible, and this flexibility ensures full employment of all these factors. All markets are competitive, and the representative firm in each of these four sectors maximises profit.

We use the following notations.

$K$ = Exogenously given domestic capital endowment.

$H$ = Exogenously given foreign capital endowment.

$L$ = Exogenously given labour endowment.

$L_{U}$ = Level of unemployed labour in the urban sector.

$N$ = Exogenously given land endowment.

$a_{j i}$ = Per unit requirement of the $j$ th input in $i$ th sector for $j = L, K, N, H$ and $i = 1,2, 3,4$ .

$P_{i}$ = price of the ith sector for $i = 1,2, 3,4$ .

$w$ = Wage rate in Sectors 1, 2 and 3.

$\bar{w}$ = Fixed wage rate in Sector 4.

$r$ = Rental rate on domestic capital.

$h$ = Rental rate on foreign capital.

$R$ = Rental rate on land.

$X_{i}$ = Level of output of $i$ th sector for $i = 1,2, 3,4$ .

$t$ = Transportation cost from rural to the urban region for migrated labour.

$θ_{j i}$ = Distributive share of $j$ th input in $i$ th sector for $j = L, K, N, H$ and $i = 1,2, 3,4$ .

$λ_{j i}$ = Proportion of $j$ th input employed in $i$ th sector for $j = L, K, N, H$ and $i = 1,2, 3,4$ .

$S_{j i}^{h}$ = The elasticity of factor output coefficient of $j$ th factor in $h$ th sector with respect to the price of the ith factor for $i$ , $j$ = $L, K$ and $h$ =1, $2$ .

For example, $S_{L N}^{1} = (\frac{R}{a_{L 1}}) (\frac{{\partial a}_{L 1}}{\partial R}), S_{L L}^{1} = (\frac{w}{a_{L 1}}) (\frac{{\partial a}_{L 1}}{\partial w})$ and so on. $S_{j i}^{h} > 0 f o r j \neq i$ ; and $S_{j j}^{h} < 0$ .

Also, $S_{j i}^{h} + S_{j j}^{h} = 0 f o r j \neq i$ .³

$\hat{x}$ = $\frac{d x}{x}$ = Relative change in $x$ .

The following equations describe the model.

a_{L 1} w + a_{N 1} R = P_{1}

(1)

a_{L 2} w + a_{K 2} r + a_{N 2} R = P_{2}

(2)

a_{L 3} w = P_{3}

(3)

a_{L 4} \bar{w} + a_{h 4} h = P_{4}

(4)

D_{2} (P_{2}) = X_{2}

(5)

D_{3} (P_{3}) = X_{3}

(6)

a_{L 1} X_{1} + a_{L 2} X_{2} + a_{L 3} X_{3} + a_{L 4} X_{4} + L_{U} = L

(7)

a_{N 1} X_{1} + a_{N 2} X_{2} = N

(8)

a_{K 2} X_{2} = K

(9)

a_{h 4} X_{4} = H

(10)

\frac{a_{L 4} X_{4}}{a_{L 4} X_{4} + L_{U}} (\bar{w} - t) = w

(11)

Equations (1)–(4) represent competitive equilibrium conditions in Sectors 1–4. In each of these three equations, the right-hand side represents the marginal cost of production. Marginal cost is equal to the average cost because the production function satisfies constant returns to scale. Price equals marginal cost in the profit-maximising equilibrium of a competitive firm, and all firms are identical in each of these two sectors. Sectors 1 and 2 form a Heckscher–Ohlin nugget or subsystem. Factor-output coefficients, $a_{j i},$ are also functions of factor price ratios because the iso-quant in each of these sectors is smooth, convex and the representative competitive firm maximises profit. Equations (5) and (6) represent the demand–supply equality condition in the market for non-traded sectors. The left-hand side of Equations (5) and (6) represents the demand function for two non-farm goods, respectively. Demand for non-traded goods varies inversely with their price. Equations (7)–(10) stand for equilibrium conditions in the labour market, land market, domestic capital market and foreign capital market, respectively. $a_{L 1} X_{1} + a_{L 2} X_{2} + a_{L 3} X_{3} + a_{L 4} X_{4}$ represents demand for labour. $L_{U}$ is unemployed labour. The summation of these two is equal to the total given supply of labour $(a_{N 1} X_{1} + a_{N 2} X_{2})$ stands for demand for land, and supply–demand equality in the land market is attained by the flexibility of factor price, $R$ . Agricultural sector and non-farm sector pay the same factor prices for land because the land is perfectly mobile among these two sectors. $a_{K 2} X_{2}$ stands for the demand for domestic capital. Domestic capital is a specific factor, and it has its demand only from the non-farm sector type 1 (Sector 2). Return for domestic capital, $r$ , is determined by equating land demand with its supply. $a_{h 4} X_{4}$ stands for demand for foreign capital. Foreign capital is a specific factor, and it has its demand only from the urban manufacturing sector. Return for foreign capital, $h$ , is determined by equating land demand with its supply. Equation (11) implies the Harris–Todaro migration equilibrium condition. Here, $a_{L 4} X_{4}$ is the total employment of labour in Sector 4, $(a_{L 4} X_{4} + L_{U})$ is the total labour supply in Sector 4 and $\frac{a_{L 4} X_{4}}{a_{L 4} X_{4} + L_{U}}$ is the probability of getting a job in Sector 4. As Sector 4 is located in an urban region and far from a rural region, people who are migrating from rural to urban areas also need to pay a transportation cost. The difference between the fixed wage rate in Sector 4 and the transportation cost is the actual earning of the labour in urban region. Finally, $\frac{a_{L 4} X_{4}}{a_{L 4} X_{4} + L_{U}} (\bar{w} - t)$ is expected earnings from Sector 4. This must be equal to the rural wage rate in equilibrium. Until these two sides become equal, migration of labour continues.

As defined earlier, a village is transformed into a CT when 75% of its male main workforce works in the non-farm sector, and its population exceeds 5,000, and the population density exceeds 400/km². Let us assume that the size and density of the population living in a village fulfil the criteria of a CT, but the labour market condition is not fulfilled. The percentage of its male workforce that gets employed in the village itself depends on the labour market conditions prevailing in the village and the adjacent city. The present paper focuses on this aspect of the formation of a CT. For facilitating the analysis, we assume all working members in the rural households are male.

In this theory, we have included two rural non-farm sectors, Sector 2 (non-farm sector type 1) and Sector 3 (non-farm sector type 2) and differentiated them in terms of input usage. Sector 2 uses land, labour and domestic capital, similar to the construction sector, which has experienced a boom in the rural sector with expansion in rural road infrastructure after the year 2,000. Sector 3, on the other hand, is like rural transport, whose main input is labour, though a small amount of capital may be used there. However, for simplicity, we assume it to be zero. Hence, the condition for CT dynamics is contraction of Sector 1 or expansion of Sectors 2 and 3 in aggregate, or both simultaneously, so that the following labour market condition is to be satisfied at a given point of time for a village to be reclassified as CT, which satisfies the other two conditions of population of 5,000 or more and population density of 400/km².

a_{L 2} X_{2} + a_{L 3} X_{3} \geq \frac{3}{4} {(a}_{L 1} X_{1} + a_{L 2} X_{2} + a_{L 3} X_{3})

Working of the Model

There are 11 unknowns in the model: $w$ , $r$ , $R$ , $h$ , $P_{2}$ , $P_{3}$ , $X_{1}$ , $X_{2}$ , $X_{3}$ , $X_{4}$ and $L_{U}$ . So, we have 11 independent equations with 11 unknowns. $H$ , $t$ , $K$ and $L$ four parameters in this model. In the price system, there are four equations with six unknowns namely, $w$ , $r$ , $R$ , $h$ , $P_{2}$ and $P_{3}$ . So the equational structure does not satisfy the decomposable property.

Equation (4) determines $h$ . Equation (3) solves for $w$ in terms of $P_{3}$ . Equation (1) determines $R$ in terms of $P_{3}$ . Equation (2) solves for $r$ in terms of $P_{2}$ and $P_{3}$ . Equation (10) solves for $X_{4}$ . Equation (11) determines $L_{U}$ in terms of $P_{3}$ . Equation (9) solves for $X_{2}$ in terms of $P_{2}$ and $P_{3}$ . Equation (8) determines $X_{1}$ in terms of $P_{2}$ and $P_{3}$ . Equation (7) determines $X_{3}$ in terms of $P_{2}$ and $P_{3}$ . Finally, Equation (5) and (6) simultaneously solves for $P_{2}$ and $P_{3}$ .

Initially, we examine the impact of a change in foreign capital on different variables of the model, keeping other parameters fixed; that is, $\hat{H} \neq 0$ but $\hat{t} = \hat{K} = \hat{L} = 0$ .

From Equations (1)–(3), (5) and (9), we get⁴

A {\hat{P}}_{2} = B {\hat{P}}_{3}

(12)

A = (e_{P_{2}} + \frac{S_{K K}^{2}}{θ_{K 2}})

B = \{- \frac{S_{K K}^{2}}{θ_{K 2}} (\frac{θ_{N 2} θ_{L 1} - θ_{L 2} θ_{N 1}}{θ_{N 1}}) + \frac{S_{K N}^{2} θ_{L 1}}{θ_{N 1}} - S_{K L}^{2}\} .

Here, $e_{P_{2}} < 0$ is the price elasticity of demand for Sector 2. We assume $A < 0$ to ensure Walrasian stability in the market of the non-traded Sector 2.

$(θ_{N 2} θ_{L 1} - θ_{L 2} θ_{N 1})$ represents the factor-intensity ranking between Sectors 1 and 2. If Sector 1 is land-intensive compared to Sector 2, then $(θ_{N 2} θ_{L 1} - θ_{L 2} θ_{N 1}) < 0$ . As, $S_{K K}^{2} < 0$ ; the first expression of $B$ is negative. As, $S_{K N}^{2}$ and $S_{K L}^{2}$ is positive, so the second and third expressions are positive and negative, respectively. Now if $S_{K N}^{2}$ is very small, that is, the elasticity of $a_{K 2}$ with respect to the return to land is very small then $B$ is negative.

From Equations (1)–(3), (8) and (12), we get⁴

{\hat{X}}_{1} = C {\hat{P}}_{3}

(13)

where

C = \frac{1}{λ_{N 1}} [- λ_{N 2} [- \frac{S_{K K}^{2}}{θ_{K 2}} \frac{B}{A} + \{- \frac{S_{K K}^{2}}{θ_{K 2}} (\frac{θ_{N 2} θ_{L 1} - θ_{L 2} θ_{N 1}}{θ_{N 1}}) + \frac{S_{K N}^{2} θ_{L 1}}{θ_{N 1}} - S_{K L}^{2}\} \frac{1}{θ_{L 3}}]

+ (λ_{N 1} S_{N N}^{1} + λ_{N 2} S_{N N}^{2}) \frac{θ_{L 1}}{θ_{L 3} θ_{N 1}} - (λ_{N 1} S_{N L}^{1} + λ_{N 2} S_{N L}^{2}) \frac{1}{θ_{L 3}}

- λ_{N 2} S_{N K}^{2} \{\frac{1}{θ_{K 2}} \frac{B}{A} + \frac{1}{θ_{L 3} θ_{K 2}} (\frac{θ_{N 2} θ_{L 1} - θ_{L 2} θ_{N 1}}{θ_{N 1}})\}] .

The first and fourth expressions of $C$ are positive and the second and third expressions are negative. If $λ_{N 2}$ is very small, then $C$ is negative. As Sector 2 labour intensive so it is obvious that the share of land in Sector 2 is less. So, from Equation (13), there is a direct relationship between the output of Sector 1 and the price of Sector 3.

From Equation (10), we get

{\hat{X}}_{4} = \hat{H}

(14)

From Equations (1), (11) and (14), we have⁴

{\hat{L}}_{U} = \hat{H} - \frac{a_{L 4} X_{4} + L_{U}}{L_{U}} \frac{{\hat{P}}_{3}}{θ_{L 3}}

(15)

Now from Equations (1)–(3), (6), (7), (12)–(15), we get⁴

{\hat{X}}_{3} = - \frac{e_{P_{3}}}{I} \frac{λ_{L 4} + \frac{L_{U}}{L}}{λ_{L 3}} \hat{H}

(16)

where

I = (e_{P_{3}} - G)

G = \frac{1}{λ_{L 3}} [- λ_{L 1} C - λ_{L 2} [- \frac{S_{K K}^{2}}{θ_{K 2}} \frac{B}{A} + \{- \frac{S_{K K}^{2}}{θ_{K 2}} (\frac{θ_{N 2} θ_{L 1} - θ_{L 2} θ_{N 1}}{θ_{N 1}}) + \frac{S_{K N}^{2} θ_{L 1}}{θ_{N 1}} - S_{K L}^{2}\} \frac{1}{θ_{L 3}}]

+ \frac{a_{L 4} X_{4} + L_{U}}{L} \frac{1}{θ_{L 3}} + E - F \frac{B}{A}];

F = \frac{{λ_{L 2} S}_{L K}^{2}}{θ_{K 2}} > 0;

E = \{- (λ_{L 1} S_{L L}^{1} + λ_{L 2} S_{L L}^{2} + λ_{L 3} S_{L L}^{3}) + (λ_{L 1} S_{L N}^{1} + λ_{L 2} S_{L N}^{2}) \frac{θ_{L 1}}{θ_{N 1}} - \frac{λ_{L 2} S_{L K}^{2}}{θ_{K 2}} (\frac{θ_{N 2} θ_{L 1} - θ_{L 2} θ_{N 1}}{θ_{N 1}})\} \frac{1}{θ_{L 3}} .

Here, $e_{P_{3}} < 0$ is the price elasticity of demand for Sector 3. We assume $I < 0$ to ensure Walrasian stability in the market of the non-traded Sector 3. So, from Equation (16), due to a rise in foreign capital, output in Sector 3 falls.

Finally, using Equations (15) and (16), we have

{\hat{L}}_{U} = \{1 + \frac{(a_{L 4} X_{4} + L_{U})}{L_{U}} \frac{e_{P_{3}}}{θ_{L 3}} \frac{\frac{λ_{L 4} + \frac{L_{U}}{L}}{λ_{L 3}}}{I}\} \hat{H}

(17)

From Equation (16), we have a positive relationship between foreign capital and unemployment of labour.

From Equations (1)–(3), (6), (9), (12) and (16), we have⁴

{\hat{X}}_{2} = - (\frac{- S_{K K}^{2}}{e_{P_{2}} θ_{K 2} + S_{K K}^{2}} + \frac{1}{θ_{L 3}}) \frac{B}{I} \frac{λ_{L 4} + \frac{L_{U}}{L}}{λ_{L 3}} \hat{H}

(18)

In the bracket, the first expression is negative and the second is positive. If, $θ_{L 3}$ is very large, then the expression in the bracket is negative. And, due to the rise in foreign capital, the output of Sector 2 expands.

From Equations (6), (13) and (16), we have

{\hat{X}}_{1} = - \frac{C}{I} \frac{λ_{L 4} + \frac{L_{U}}{L}}{λ_{L 3}} \hat{H}

(19)

From Equation (19), due to a rise in foreign capital, the output of Sector 1 contracts.

Equations (12)–(16) are summarised in the following proposition.

Proposition 1. Due to a rise in foreign capital, Sectors 1 and 3 contract, but other sectors; that is, Sectors 2 and 4 expand. Finally, a rise in foreign capital raises unemployment. If Sectors 2 and 3 expand in aggregate given the contraction of Sector 1, CT dynamics accelerate.

Let us now describe Proposition 1 intuitively. Inflow of foreign capital expands the sector as foreign capital is specific to Sector 4. As Sector 4 expands, employment rises in Sector 4. So now, less labour is available for the remaining sectors. Labour is a specific factor in Sector 3. So, Sector 3 contracts. This releases fresh labour for Sectors 2 and 1. Now, Sector 2 is labour-intensive compared to Sector 1. So, Sector 2 expands and Sector 1 contracts according to the Rybczynski theorem.⁵ As Sector 4 expands, more labour will be attracted to this sector. However, this sector cannot absorb all labour, thus, unemployment is going to rise.

Out of two rural non-farm sectors (Sectors 2 and 3), one sector (Sector 3) contracts and one sector (Sector 2) expands. Hence, the effect of capital inflow on CT dynamics is ambiguous. However, if the two rural non-farm sectors in aggregate expand, given the contraction of the rural farm sector, the CT dynamics accelerate.

Next, we analyse the impact of a change in transportation cost, keeping the other parameters fixed; that is, $\hat{t} \neq 0$ but $\hat{H} = \hat{K} = \hat{L} = 0$ .

Then from Equations (1)–(9) and (11), we have⁴

{\hat{X}}_{3} = \frac{e_{P_{3}}}{I} \frac{\frac{t}{\bar{w} - t} \frac{a_{L 4} X_{4} + L_{U}}{L_{U}}}{λ_{L 3}} \hat{t}

(20)

From Equation (7), we have a direct relationship between the output of Sector 3 and the transportation cost.

Using Equations (1)–(3) and (9), we get

{\hat{X}}_{2} = (\frac{- S_{K K}^{2}}{e_{P_{2}} θ_{K 2} + S_{K K}^{2}} + \frac{1}{θ_{L 3}}) \frac{B}{I} \frac{\frac{t}{\bar{w} - t} \frac{a_{L 4} X_{4} + L_{U}}{L_{U}}}{λ_{L 3}} \hat{t}

(21)

Using Equations (6) and (13), we have

{\hat{X}}_{1} = \frac{C}{I} \frac{\frac{t}{\bar{w} - t} \frac{a_{L 4} X_{4} + L_{U}}{L_{U}}}{λ_{L 3}} \hat{t}

(22)

Finally, using Equations (3), (6) and (11), we get⁴

{\hat{L}}_{U} = - \frac{a_{L 4} X_{4} + L_{U}}{a_{L 4} X_{4}} (\frac{1}{θ_{L 3}} \frac{1}{I} \frac{\frac{a_{L 4} X_{4} + L_{U}}{L_{U}}}{λ_{L 3}} + 1) \frac{t}{\bar{w} - t} \hat{t}

(23)

In Equation (22), the first expression in the bracket is negative and the second expression is positive. The second expression denotes the direct impact of transportation costs on unemployment. The first expression denotes the effect of transportation cost on unemployment through a change in wage rate. The effect of a change in transportation cost is summarised in the following proposition.

Proposition 2. Due to a rise in transportation cost, Sectors 1 and 3 expand, but other sectors, that is, Sectors 2 and 4, contract. The effect on CT dynamics is ambiguous.

Let us now describe Proposition 2 intuitively. A rise in transportation costs directly makes urban jobs unfavourable. So, unemployment falls initially through the direct effect. This makes more labour available for the remaining sectors. So, sector three expands where labour is a specific factor. As Sector 3 expands, there is a shortage of labour for Sectors 2 and 1. As Sector 2 is labour-intensive, Sector 2 contracts and Sector 1 expands according to the Rybczynski theorem. Due to the initial fall in unemployment, there is an excess supply of labour, which makes rural wages fall. This fall in rural wages forces labour to move from rural to urban areas, and unemployment rises indirectly. So, the final effect on unemployment is ambiguous. Since Sector 1 expands, Sector 2 contracts and Sector 3 expands, it is difficult to say whether CT dynamics accelerate. If Sectors 2 and 3 expand in aggregate more (less) than proportionally compared to Sector 1, CT dynamics accelerate (decelerate).

III. Empirical Observation

The objective of the above model is to observe how the inflow of foreign capital and changes in the transport cost led to the growth of the CT.⁶ We suggest the following reduced-form regression equation in order to test the validity of the propositions made above

y = β_{0} + β_{1} x_{1} + β_{2} x_{2} + β_{3} x_{3} + β_{23} x_{2} x_{3} + D_{d} + u

(24)

Here the dependent variable $(y)$ is a binary variable for CT defined as follows: those urban bodies which were identified by RGI as ‘would be CT for 2001’ and converted to CT in 2011 take the value 1, whereas the urban bodies which were identified by RGI as ‘would be CT for 2001’ but not reclassified as CTs in 2011 take the value 0. The independent variables are: (a) inflow of foreign capital $(x_{1})$ , (b) nearest city/ST road distance (in km) from CT $(x_{2})$ , (c) nearest highway road distance (in km) from CT $(x_{3})$ and (d) the interaction between these two distances. $D_{d}$ represents district-specific dummies, which take into account fixed effects associated with locations of CTs/villages taken up for the study, and u represents the random error associated with the regression, which we assume to be independently normally distributed. The fixed effects would usually stand for the historical or social factors that are specific to a district or zone and may have contributed to the growth of CTs. The variables that are present in the theoretical model, namely foreign capital inflow and transport cost, are replaced by proxy variables. Foreign capital is proxied by the number of banks in the nearest city/ST, and transport cost is proxied by the nearest city road distance (in km) from CT. We consider a second type of distance, which is the highway road distance (in km) from CT (not considered in the theory). We have taken two distances to have an insight into the role of transport cost on the agglomeration effect due to juxtaposition near the city. It is anticipated that the number of banks in the nearest city/ST will have a positive effect on the dependent variable. This expectation stems from the fact that, from the theoretical model, we observe that as the foreign capital inflow increases, the urban manufacturing sector expands, which means that the financial sector in urban areas expands, and as a result, the banking sector also expands, which is reflected by the expansion of a number of banks in that area. On the other hand, the distance of ‘would-be CT’ from the nearest city/highway is expected to have an ambiguous effect. An increase in transport cost not only diminishes the likelihood of commuting to nearby urban centres for non-farm job opportunities but also changes the rural labour dynamics, thereby leaving ambiguity on the probability of a ‘would-be CT’ being classified as CT. Hence, the sign of this explanatory variable is expected to be positive/negative.

Data and Estimation Procedure

The data used to analyse the above empirical model are the Census data of 2001 and 2011. These data provide a complete description of the number of villages, STs and CTs in West Bengal. We have extracted the necessary data to run the regression for the present paper. During the period 2001–2011, there was a maximum increase in the number of CTs for West Bengal among the states in India. Hence, this state becomes the focus of empirical analysis for the theoretical model set above. We processed the census data of West Bengal to construct the dataset on CTs required for the study. Table 1 provides a summary of the data. In the first row, it is observed that around 80% of the ‘would-be CT for 2011’ in West Bengal actually converted to CT in 2011. On average, the proximity to the nearest highway is 3.16 km, while the average distance to the closest city with a population exceeding one lakh is 28.86 km. Notably, the average number of banks in the nearest city is 185; however, when Kolkata is excluded from the data set, the number of observations decreases to 538, and the average number reduces to 23.

Table 1.

Summary Statistics.

	No of Observation	Mean	SD	Min	Max
Census towns (2011)	654	0.80	0.41	0	1
Distance (in km) from the nearest highway	654	3.16	3.52	0	17.23
Distance (in km) from the city	654	28.86	29.99	0	165
Number of banks in the nearby city	645	185	366.98	1	1,007
Number of banks in the nearby city (except Kolkata)	538	23	21.35	1	73

Source: All the variables are taken from the Census of India (2011).

Results

Conventional methods like Logit or Probit could be preferred to estimate the regression equation mentioned above. However, due to the presence of heteroskedastic errors, we opt to present the model using the heteroskedastic robust OLS method (Linear Probability Model) as a baseline model, acknowledging the limitations of OLS in this context. Table 2 outlines the estimated coefficients derived from the OLS estimation of Equation (19). Additionally, Tables 3 and 4 provide the results of Logit and Probit estimations of the same equation, respectively, demonstrating that the coefficients’ magnitude and direction closely resemble those obtained from OLS.

Table 2 describes the OLS estimates. Columns 1–3 correspond to all CTs, while columns 4–6 represent all CTs excluding those with Kolkata as their nearest city. The rationale for this exclusion is straightforward: Kolkata’s classification as a metropolitan city suggests it may act as an outlier within the sample. For almost all specifications, we see that the distance from the nearest highway (in km) is negative and significant at the 1% level of significance. The increase in the number of CTs indirectly corresponds to the expansion of the non-farm sector (Sectors 2 and 3 in our theoretical model). From Proposition 1, we observe that as foreign capital inflow increases, sector two expands. This is observed in the data from the fact that for every 100 increase in the number of banks (proxy of foreign capital), the probability that a would-be CT in 2001 becomes a CT in 2011 increases by 6%, which confirms the theoretical finding in the previous section. From Proposition 2, we can observe that as the transportation cost (which is proxied by the distance from the nearest city/highway) increases, the non-farm sector may expand (shrink) as a whole, which means the probability of would-be CTs actually being converted to CTs rises(reduces). In column 2, where the distance from the nearest city is used, we observed that the coefficient is positive and significant. An increase in the distance by 1 km results in a mere 0.5% rise in the likelihood of potential CTs being converted to CTs, which is considered negligible. This slight increase is attributed to the phenomenon where CTs located closer to the city have a higher chance of transitioning to municipal status due to their proximity (which we call the juxtaposition or agglomeration effect), thereby reducing the probability of becoming CTs. In column 1, we can see that for every kilometre increase in distance from the highway, the probability that a would-be CT in 2001 becomes a CT in 2011 reduces by 4.1%. In column 3, an additional interaction term is introduced alongside the two distance variables. It is observed that the coefficient of this interaction term is positive and statistically significant. This indicates that as the distance from the highway increases, the probability of a would-be CT in 2001 converting to a CT in 2011 slightly rises (by 0.04%) for CTs farther from the city compared to those nearby. This suggests a complementary relationship between these two distances. Columns 4 through 6 replicate the analysis, excluding CTs whose nearest city is Kolkata. We can observe that the significance of a number of banks disappears due to this omission. This outcome is expected because Kolkata plays a significant role in terms of attracting foreign capital inflow. Note that the significance of the distance factors remains unchanged. The last two rows of the table depict the goodness of fit for all specifications. In the last row, we can see that the prediction using this OLS model is quite impressive, more than 80% are correctly predicted.

Table 2.

Ordinary Least Squares (OLS) Estimation.

Dependent Variable: Whether a would be Census Town (CT) in 2001 Was Converted to CT in 2011 (Yes = 1, No = 0)
	All CTs of WB			Except Kolkata
	(1)	(2)	(3)	(4)	(5)	(6)
Distance from highway (in km)	−0.041*** (0.006)		−0.053*** (0.005)		−0.047*** (0.006)	−0.054*** (0.007)
No of banks (’00)	0.060*** (0.013)	0.064*** (0.014)	0.056*** (0.014)	0.101 (0.130)	0.164 (0.108)	0.098 (0.112)
Distance from city (in km)		0.005*** (0.001)	0.003*** (0.001)	0.005*** (0.001)		0.004*** (0.0007)
Distance from highway (in km) × Distance from city (in km)			0.0004*** (0.0007)			0.0003* (0.0001)
Constant	0.801*** (0.054)	0.545*** (0.062)	0.715*** (0.058)	0.542*** (0.067)	0.797*** (0.059)	0.711*** (0.061)
Observations	645	645	645	538	538	538
R ²	0.218	0.177	0.342	0.214	0.283	0.346
Percentage correctly predicted	84.10	82.57	86.39	82.57	84.71	86.70

Notes: Coefficients of the regression Equation (19) are presented in table. Robust standard errors are shown in the parentheses. For all specifications, we have used district fixed effects to remove the unobserved heterogeneity due to different locations or backgrounds of the districts. Columns 1–3 correspond to all CTs, while columns 4–6 represent all CTs excluding those with Kolkata as their nearest city. *, and *** indicate that coefficients are significant at 10%, and 1% levels of significance, respectively.

Robustness

To substantiate observations made earlier, we also estimate the regression Equation (19) using the Logit and Probit models, which are represented by Tables 3 and 4, respectively. What we observe here is similar to that observed in the OLS model. We report the margins of the coefficients of the Logit and Probit estimation. To measure the goodness of fit of the models, we report pseudo-R-squared and the percentage of correct predictions. Both measures indicate that the models are well fitted.

Table 3.

Logit Model.

Dependent Variable: Whether a would be Census Town (CT) in 2001 Was Converted to a CT in 2011 (Yes = 1, No = 0)
	All CTs of WB			Except Kolkata
	(1)	(2)	(3)	(4)	(5)	(6)
Distance from highway (in km)	−0.032*** (0.004)		−0.033*** (0.005)		−0.037*** (0.004)	−0.035*** (0.006)
No of Banks (’00)	0.051*** (0.011)	0.055*** (0.012)	0.051*** (0.010)	0.077 (0.106)	0.158 (0.100)	0.049 (0.092)
Distance from city (in km)		0.006*** (0.001)	0.006*** (0.001)	0.006*** (0.001)		0.006*** (0.001)
Distance from highway (in km) × Distance from city (in km)			0.00008 (0.0002)			−0.00001 (0.0002)
Observations	627	627	627	505	505	505
Pseudo R²	0.193	0.181	0.298	0.135	0.183	0.286
Percentage correctly predicted	84.71	81.96	86.39	79.97	83.49	84.25

Notes: Margins of the coefficient from Logit regression are shown (margins are defined as the $\frac{\partial P (y = 1| x)}{\partial x_{i}} = g (β_{0} + x β) β_{i}$ where $g (z) \equiv \frac{d G (z)}{d z}, z = β_{0} + x β$ and $G (z) = \frac{\exp (z)}{1 + \exp (z)}$ . Standard errors are shown in parentheses. For all specifications, we have used district fixed effects to remove the unobserved heterogeneity due to different locations or backgrounds of the districts. Columns 1–3 correspond to all CTs, while columns 4–6 represent all CTs excluding those with Kolkata as their nearest city. *** tells that coefficients are significant at 1% levels of significance.

Table 4.

Probit Estimation.

Dependent Variable: Whether a would be Census Town (CT) in 2001 Converted to a CT in 2011 (Yes = 1, No = 0)
	All CTs of WB			Except Kolkata
	(1)	(2)	(3)	(4)	(5)	(6)
Highway road distance (in km)	−0.033*** (0.004)		−0.035*** (0.006)		−0.037*** (0.004)	−0.036*** (0.006)
No of banks (’00)	0.044*** (0.009)	0.045*** (0.009)	0.041*** (0.008)	0.061 (0.100)	0.167* (0.099)	0.062 (0.089)
Distance from city (in km)		0.006*** (0.001)	0.006*** (0.001)	0.006*** (0.001)		0.006*** (0.001)
Distance from highway (in km) × Distance from city (in km)			0.0001 (0.0002)			−0.00003 (0.0002)
Observations	627	627	627	505	505	505
Pseudo R²	0.185	0.171	0.290	0.133	0.181	0.285
Percentage correctly predicted	84.10	81.80	85.47	79.66	82.87	83.94

Notes: Margins of the coefficient from the Probit regression are shown (margins are defined as the $\frac{\partial P (y = 1| x)}{\partial x_{i}} = g (β_{0} + x β) β_{i}$ where $g (z) \equiv \frac{d G (z)}{d z}, z = β_{0} + x β$ and $G (z) = Φ (z) \equiv \int_{- \infty}^{z} ϕ (v) d v,$ where $ϕ (z)$ is the standard normal density. Standard errors are shown in parentheses. For all specifications, we have used district fixed effects to remove the unobserved heterogeneity due to different locations or backgrounds of the districts. Columns 1–3 correspond to all CTs, while columns 4–6 represent all CTs excluding those with Kolkata as their nearest city. *, and *** tells that coefficients are significant at 10%, and 1% levels of significance respectively.

IV. Conclusion

Although the astonishing growth of CTs in India has been in the limelight in the context of urbanisation in recent times, there is a dearth of literature addressing this issue. The paper constructs a theoretical model with empirical support to check how factors like the inflow of foreign capital and the cost of transport to the nearby large urban area affect the formation of the CTs. As the theory is tested empirically by using data from the state of West Bengal during 2001–2011, where the growth rate of CTs had been one of the highest in India, it turns out that the inflow of foreign capital is not a major factor explaining the birth of CTs if we exclude CTs that have grown around Kolkata. However, transport costs play a vital role in CT formation. The paper finds that CTs grow around existing urban centres and away from them, due to changes in urban as well as rural labour dynamics, with changes in transport costs. The interesting finding of the paper is that where the rise in distance and hence transport cost from the nearest city raises the chance of a village becoming a CT, the rise in the distance from the nearest highway does the opposite. The nearness to the city actually strengthens the force of agglomeration towards the city, which we refer to as the ‘juxtaposition effect’, whereas the distance from a nearby highway is crucial for the burgeoning of the rural non-farm sector. All these findings are significant and novel in the context of urban economic literature in India.

The theory and empirics developed in this paper have some limitations. First, the theory precludes the role of city-specific factors like urban rent, employment opportunity in the urban informal sector and agricultural distress in CT dynamics. Second, on the empirical side, using urban banks as a proxy for the inflow of foreign capital is not enough. However, due to a lack of data as well as the expansion of the financial sector as a consequence of formal sector expansion, foreign capital inflow compels us to think in this way. The treatment of ‘transport cost’ in this chapter is not adequate. The distance from the nearest urban centre may not be a good measure of transport cost since the quality of roads and transport services matters. Thus, we included the distance from the nearest highway to corroborate our results. Despite the limitations, this paper is a unique contribution to urban economics since the exploitation of the general equilibrium framework with empirical testing will be helpful for future research in this area.

Footnotes

Acknowledgement

The authors thank the anonymous referees for their valuable comments and suggestions to improve the paper.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The authors received no financial support for the research, authorship and/or publication of this article.

Notes

ORCID iDs

Priya Brata Dutta

Saumyabrata Chakrabarti

Imdadul Islam Halder

Appendix A

References

Aggarwal

(2018). Do rural roads create pathways out of poverty? Evidence from India. Journal of Development Economics , 133, 375–395.

Balakrishnan

(2013). Highway urbanization and land conflicts: The challenges to decentralization in India. Pacific Affairs , 86(4), 785–811.

Baum-Snow

, Henderson

J. V.

, Turner

M. A.

, Zhang

, & Brandt

(2020). Does investment in national highways help or hurt hinterland city growth? Journal of Urban Economics , 115, 103124.

Basak

, & Gupta

(2024). Skilled emigration in a world of variety. Research in Economics , 78(3), 1–11.

Bhide

(2013). The invisible migrant. Urban Transition. Info Change Agenda , 10–13.

Census of India. (various years). Town and village directory . Office of the Registrar General and Census Commissioner, Government of India.

Chandra

, & Thompson

(2000). Does public infrastructure affect economic activity? Evidence from the rural interstate highway system. Regional Science and Urban Economics , 30(4), 457–490.

Chandra

, & Khan

M. A.

(1993). Foreign investment in the presence of an informal sector. Economica , 60(237), 79–103.

Chaudhuri

(2000). Rural–urban migration, informal sector, urban unemployment and development policies: A theoretical analysis. Review of Development Economics , 4(3), 353–364.

10.

Chaudhuri

, & Mukherjee

(2002). Removal of protectionism, foreign investment and welfare in a model of informal sector. Japan and the World Economy , 14, 101–116.

11.

Chaudhuri

, & Mukhopadhyay

(2009). The Harris–Todaro migration model and introduction of the informal sector. In Revisiting the informal sector: A general equilibrium approach (pp. 33–53). Springer.

12.

Chaudhuri

, Bhattacharyya

, & Bhattacharya

(2024). Rural urban migration with heterogeneous firms, heterogeneous labourers and the effect of wage subsidy. Theoretical and Applied Economics , 28(3), 143–156.

13.

Chaudhuri

, & Yabuuchi

(2008). Foreign capital and skilled-unskilled wage inequality in a developing economy with non-traded goods. In Marjit

& Yu

(Eds.), Contemporary and emerging issues in trade theory and policy (pp. 225–250). Emerald Group Publishing.

14.

Chaudhuri

, Yabuuchi

, & Mukhopadhyay

(2006). Inflow of foreign capital and trade liberalisation in a model with an informal sector and urban unemployment. Pacific Economic Review , 11(1), 87–103.

15.

Ghani

, Goswami

A. G.

, & Kerr

W. R.

(2012). Is India’s manufacturing sector moving away from cities? (NBER Working Paper No. 17992). National Bureau of Economic Research.

16.

Guin

, & Das

D. N.

(2015). New census towns in West Bengal: “Census activism” or sectoral diversification? Economic & Political Weekly , 50(14), 68–72.

17.

Gupta

M. R.

(1993). Rural–urban migration, informal sector and development policies: A theoretical analysis. Journal of Development Economics , 41(1), 137–151.

18.

Harris

J. R.

, & Todaro

M. P.

(1970). Migration, unemployment and development: A two-sector analysis. American Economic Review , 60(1), 126–142.

19.

Helpman

(1998). The size of regions. In Pines

, Sadka

, & Zilch

(Eds.), Topics in public economics: Theoretical and applied analysis (pp. 33–54). Cambridge University Press.

20.

Khan

M. A.

(1982). Tariffs, foreign capital and immiserizing growth with urban unemployment and specific factors of production. Journal of Development Economics , 10(2), 245–256.

21.

Krugman

(1991). Increasing returns and economic geography. Journal of Political Economy , 99(3), 483–499.

22.

Krugman

, & Elizondo

R. L.

(1996). Trade policy and the third world metropolis. Journal of Development Economics , 49(1), 137–150.

23.

Kundu

(2011a). Method in madness: Urban data from 2011 census. Economic & Political Weekly , 46(40), 13–16.

24.

Kundu

(2011b). Politics and economics of urban growth. Economic & Political Weekly , 46(20), 10–12.

25.

Mahajan

, & Nagaraj

(2017). Rural construction employment boom during 2000–12: Evidence from NSSO surveys. Economic & Political Weekly , 52(52), 54–63.

26.

Marjit

(1994). On the participation of domestic capital in a foreign enclave: A general equilibrium analysis. Journal of Economic Integration , 9(4), 489–501.

27.

Marjit

, & Kar

(2015). Broader implications of labor market reforms in India: A general equilibrium perspective. Indian Economic Review , 49(1), 27–35.

28.

Mukhopadhyay

, Zerah

M.-H.

, Samanta

, & Maria

(2016). Understanding India’s urban frontier: What is behind the emergence of census towns in India? (World Bank Policy Research Working Paper No. 7923). World Bank.

29.

National Sample Survey Office (NSSO). (various years). Reports from various rounds . Ministry of Statistics and Programme Implementation, Government of India.

30.

Pradhan

K. C.

(2013). Unacknowledged urbanization: New census towns of India. Economic & Political Weekly , 48(36), 43–51.

31.

Sharma

(2013). Exclusionary urbanization and changing migration patterns in India: Is commuting by workers a feasible alternative? Indira Gandhi Institute of Development Research.

32.

Van Duijne

R. J.

, & Nijman

(2019). India’s emergent urban formations. Annals of the American Association of Geographers , 109(6), 1978–1998.

33.

Zhu

(2017). The urban transition and beyond: Facing new challenges of the mobility and settlement transitions in Asia . Paper presented at the United Nations Expert Group Meeting on Sustainable Cities, Human Mobility and International Migration.