This paper investigates the geographic concentration of knowledge and tech nology-intensive (KTI) industries, covering 0.43 million establishments across various districts of rural and urban areas in India. Using the spatially weighted Ellison-Glaeser index, cartogram and choropleth map results show that few KTI industries are highly geographically concentrated in urban and rural areas, specific to certain districts and a few Indian states. Within highly employable states of India, workers are employed in only a particular location of a few districts. Also, we differentiate between urban and rural concentrated and urban and rural dispersed districts within highly employable states. In addition, results validate the extent of the geographical concentration of KTI industries in rural and urban areas of highly employable Indian states. Further, results exhibit that industries spatially concentrate in only a few locations across specific districts in India, indicating natural advantages and other economic forces are pretty strong in certain areas. Besides, results suggest that the demand-based networks and push-and-pull supply chains are well established in a specific location of a few districts, incentivizing other firms to locate their business, which creates a spatial spillover effect and benefits all economic agents. Empirical results suggest that policymakers in India could unleash the resource potential of spatially concentrated districts by implementing a location-based policy and considering multi-level governance and informal and formal institutions, which could further boost regional economic growth.
The uneven geographical concentration of economic activity is a feature that is pervasive in many countries (Braunerhjelm & Johansson, 2003; Vitali et al., 2013; Crafts & Klein, 2021). In the era of globalization, innovation is at the core of competitive advantage attainment and altering the geographic landscape as knowledge clusters become the significant drivers of the prosperity of nations (Huggins, 2008). The idea of geographic concentration goes way back to Marshall (1890), arguing that the agglomeration of firms affects productivity through sharing intermediate inputs at low cost, a local labour market pooling, and the exchange of ideas between firms. Knowledge spillovers play an essential role in understanding the uneven growth patterns of regions (Glaeser et al., 1992; Henderson et al., 2001). A firm’s survival depends on cutting-edge knowledge, due to which firms are dispersing their R&D activities to tap into geographic centres of excellence worldwide (Lundvall, 2007; Mudambi, 2008). It implies the importance of geographical proximity in creating and diffusing knowledge as it requires face-to-face interaction between different economic actors (Becker et al., 1999; Breschi & Lissoni, 2009; Balland et al., 2015). Under this backdrop of geographical proximity, the pivotal question is whether Knowledge and Technology-Intensive (KTI) industries spatially concentrate in rural and urban areas in India. More specifically, the question arises in the context of emerging economies like India, how KTI industries are geographically concentrated or dispersed to a different location, and whether the KTI industries have any role in regional economic growth for hiring a massive surplus labour force in India. This paper investigates this central question in the context of KTI industries in India.
The artificial intelligence revolution and innovative development have entirely transformed the global economy (Burström et al., 2021; Korinek & Stiglitz, 2021). Globally, KTI industries invest the most significant shares of their output in research and development (R&D)—contributing 11% to both US gross domestic product (GDP) ($2.3 trillion) and global GDP ($9.2 trillion) in 2019 (National Science Foundation, 2022)1. KTI industry’s value-added output has more than doubled from $3.4 trillion in 2002 to $9 trillion in 2018. Out of global KTI output ($9 trillion) in 2018, 64% came from medium-high R&D intensive industries, while the high R&D intensive industry’s output contribution is only 36% (National Science Foundation, 2020). In the case of India, value-added by KTI industries increased from $34.5 billion in 2003 to $209.2 billion in 20182. Within Asia, India holds the fourth position in value addition of KTI industries, after China, Japan, and South Korea in 2018. The United States (US) and China share the top spot as the world’s largest producers of total KTI output (each with a 25% global share of KTI value added in 2019) (National Science Foundation, 2020).
The previous studies show that most Indian agglomeration literature focuses on the IT cluster and its development. However, the Indian literature completely neglects the KTI industries and their contribution to regional economic growth. The KTI industries include IT and software publishing, scientific research and development, air and spacecraft, pharmaceuticals, computer, electronic and optical products, motor vehicles, medical and dental instruments, railroad, chemicals, and electrical industries. The United States and China invest heavily in the research and commercialization of artificial intelligence (AI) technologies (National Science Foundation, 2020). However, the US leads in KTI services industries, whereas China leads KTI manufacturing industries production. Moreover, US KTI production is geographically concentrated, with 15 states accounting for 76% of the total value added domestically by KTI industries. KTI manufacturing in the US is concentrated in the Midwest, along the coasts, and in a few states in the South, while KTI services are concentrated along the coasts and a few Southwestern states (National Science Foundation, 2022). US KTI industries are the principal force behind the nation’s research and development (R&D) enterprises. From an economic standpoint, R&D activities constitute a critical component of US economic growth and competitiveness.
India is one of the world’s largest functioning democracies and has a critical mass of English-speaking knowledge workers and free-market economic institutions. Besides, India has developed a broad and diversified Information and Communication Technology (ICT) infrastructure in recent years, reflecting its advantageous position. In his famous book “The World Is Flat: A Brief History of the Twenty-First Century”, Thomas Friedman wrote that the economic playing field was levelling out for India. The best example is Bangalore, India’s Silicon Valley, which comprises modern IT infrastructure and technologies backed by the top technical institute and has India’s most competent engineers. The Bangalore IT cluster is the fastest-growing software cluster outside the USA (Parthasarathy, 2004; Huggins, 2008). Various studies (Parthasarathy, 2004; Khomiakova, 2007; Grondeau, 2007; Lorenzen & Mudambi, 2013) focus on the IT cluster formation and the importance of social embeddedness in decentralized network structures.
Moreover, specific to Indian industries, only a few previous studies (Lall et al., 2004; Lall & Chakravorty, 2005; Fernandes & Sharma, 2012) measure the spatial concentration of Indian manufacturing industries in their empirical studies. Mukim’s (2015) study finds that buyer-seller linkages and technology spillovers are the most significant factors in explaining the co-agglomeration of formal-informal manufacturing enterprises. Desmet et al. (2015) study found that agglomeration forces in service sectors still dominate dispersion forces in high-density areas, given the role played by ICT. Ghani et al. (2016) studied the spatial pattern of manufacturing and service industries in India from 2001 to 2010. Their study finds that the organized manufacturing sector moves away from urban cores to the rural periphery while services move towards the urban centres. According to the 2011 Economic Census, 72.4% of the workforce and 68.8% of the country’s population exist in rural areas of India (Chand et al., 2017). Also, the rural-urban divide accounts for a large share of spatial inequality with significant differences in output per capita and access to core public services, such as electricity, roads, and education in India (Joumard et al., 2017)3. Regional inequality is of interest not only for equity reasons but also for economic development (Achten & Lessmann, 2020).
Nevertheless, India’s gross domestic expenditure on R&D (GERD) increased from $19180.50 million in 2000 to $57825.12 million in 2018. However, India’s GERD as a percentage of GDP decreased from 0.77% in 2000 to 0.66% in 20184. It indicates ample opportunities for India to invest in R&D activities to enhance KTI industries’ output significantly. To boost economic output and exports of products and services in the foreign market, the competitiveness and productivity of firms matter. For regional economic growth and productivity of firms, the geographic concentration of economic activities matters (Combes & Gobillon, 2015; Graham et al., 2010). This is a prerequisite to exploring the geographical concentration of KTI industries to design appropriate urban and rural policies that attract or strengthen knowledge-intensive industries. Chen’s (2020) study reveals that in China, upgrading the industrial structure significantly affects poverty reduction in urban areas but not in rural areas. This indicates that differences between urban and rural areas must be considered while framing regional policies to upgrade the industrial structure.
Our study contributes to Indian agglomeration literature, particularly KTI industries, in several ways. First, this paper estimates the geographic concentration of KTI industries at a 3-digit National Industrial Classification (NIC) code, using Economic Census (2013) data at the establishment levels across various rural and urban areas in India5. Besides, this paper examines the spatial distribution and magnitude of spatial concentration of different KTI industries in rural and urban areas. The spatial distribution of KTI industries visualizes through the cartogram technique, and geographical concentration is measured using a spatially weighted Ellison-Glaeser index. Second, this study measures the geographical concentration of highly concentrated KTI industries within highly employable states across urban and rural India6. Third, we identify the districts where excess employment concentration prevails across the highly concentrated KTI industries. More specifically, we distinguished between urban and rural concentrated and urban and rural dispersed districts within highly employable states using a choropleth map.
The remainder of the paper is organized as follows. Section 2 provides a snapshot of previous studies of the agglomeration literature. Section 3 presents the data description and methodology for computing the geographical concentration of industries. Section 4 summarizes our empirical findings, and finally, Section 5 offers concluding remarks.
Many scholars examined the spatial concentration of knowledge-intensive industries (KTI) empirically. The idea of spatial concentration of firms goes way back to Marshall’s (1890) pioneering work. Marshall argued that three sources are essential for firms’ geographical concentration: input sharing, labour market pooling, and knowledge spillovers7. However, in this study, we specifically focus on knowledge spillovers, which refer to the transfer or flow of knowledge between firms or workers close to each other. Jacobs (1969) argues that a diversity of regional economic activity nurtures innovation and growth through inter-industry knowledge spillovers rather than intra-industry spillovers. Porter (1990) supports intra-industrial spillovers but states that competition among firms incentivizes them to exchange knowledge and innovate. The seminal work of Glaeser et al. (1992) sparked a substantial volume of empirical research to examine whether agglomeration externalities matter for knowledge creation and innovation8. Various studies (Acs et al., 1992; Jaffe et al., 1993; Zucker et al., 2002) showed that knowledge spillovers are geographically localized and significant at a local level in the context of the USA. In their study for the USA, Audretsch and Feldman (1996) observed that industries in which knowledge spillovers are predominant tend to be more geographically concentrated compared to industries where knowledge externalities are less prevalent. Moreover, Buzard et al. (2020) and Ganguli et al. (2020) research shows that geographical proximity facilitates knowledge spillovers.
Knowledge is essential to boost competitiveness and innovation performance, form local surroundings and experience spatial clustering (Malmberg et al., 1996). Tödtling et al. (2006) study reveals that knowledge intensity has played a vital role in developing a knowledge-based economy, implying the importance of knowledge-intensive industries. Paci and Usai’s (1999) study shows evidence of cross-border technological spillovers, but agglomeration effects die out with the increase in distance, implying knowledge spillovers are bounded spatially. In contrast, Li’s (2014) study reveals that border and distance effects increase over time due to the strengthening of knowledge agglomeration but decrease with the age of patents. Moreover, Andersson et al. (2016) reveal that knowledge spillovers are spatially and sensitive to geographical distance. Another dimension in the knowledge spillover literature is that excessive reliance on local knowledge can prompt territorial lock-in effects, implying the importance of inter-regional linkages. A substantial body of literature (Trippl et al., 2009; Eriksson & Lengyel, 2019; Ascani et al., 2020; Balland & Boschma, 2021) shows that Interregional linkages prevent region’s locked-in tendency. Besides, Mudambi and Swift (2012), Beugelsdijk and Mudambi (2014), and Lorenzen and Mudambi (2013) have done extensive research on the spatial dimension of FDI and modes of entry of Multinational enterprises (MNEs) into the host country’s regional location. These studies trace the origins of research on geographic clusters and identify the seminal contributions focusing on the role of MNEs, connectivity of firms of different clusters through knowledge, R&D, innovation of new technology, etc.
In the Indian context, various studies explore the appropriate determinants of firm locational choice and productivity (Behera, 2017). Similarly, Behera (Behera et al., 2012; Behera, 2015a, 2015b) find that R&D and technology import intensity enhances the productivity of Indian manufacturing industries. Lall et al. (2004) find that urbanization economies have a significant cost-reducing effect on firms that lead to industrial clustering in metropolitan areas in India. Chakravorty et al. (2005) studied eight industrial sectors in three Indian metropolises (Mumbai, Kolkata, and Chennai) to determine whether localization economies play a substantial role in cluster formation. Mukim’s (2015) study finds that buyer-seller linkages and technology spillovers are the most significant factors in explaining the co-agglomeration of formal-informal manufacturing enterprises. The literature on productivity spillover focusing on Indian industries revealed some interesting facts. Kathuria (2002) finds that local Indian firms in the manufacturing industry could reap the fruits of knowledge spillovers from foreign-owned firms only if they can decode the spilt knowledge’s technicalities. In this genre, Franco and Sasidharan (2010) study the effects of export spillovers on emerging markets, particularly India, from 1994 to 2006. They have shown that in-house research efforts are the most crucial factor for grabbing the benefits of technology spillovers than any other external sources of technology.
The research on Indian IT clusters preceding 2007 mainly focuses on Bangalore. Parthasarathy (2004) reveals the importance of social embeddedness in agglomeration by analyzing how changing state-society relations shaped the software industry in Bangalore. Khomiakova (2007) studied multiple IT clusters development in India, and Grondeau (2007) examined the characteristics of ICT clusters, mainly focusing on Bangalore and Hyderabad clusters. Lorenzen and Mudambi (2013) propose that clusters linked to the global economy through decentralized network structures have the most potential for local spillovers. Their empirical study considers IT clusters in Bangalore and the Indian film entertainment cluster (Bollywood). Besides this, various research (Kerr, 2008; Sonderegger & Täube, 2010; Zaheer et al., 2009) shows diaspora (non-local network) plays a substantial role in cluster formation.
Nevertheless, to our best understanding, none of the previous literature examines the spatial concentration of KTI industries covering 0.43 million establishments across rural and urban Indian districts. Therefore, this paper bridges the research gap in the agglomeration literature by measuring the extent of the geographical concentration of KTI industries and visualizing the spatial distribution of highly concentrated KTI industries across various Indian states. Within highly employable Indian states, our empirical interest is to quantify which KTI industries are geographically concentrated using the spatially weighted Ellison-Glaeser index. This index captures neighbourhood effects and, based on our understanding, this is the first study to measure the geographical concentration of KTI industries within highly employable Indian states in rural and urban India. Besides, this study also computes the excess employment concentration in India and tries to locate the districts where agglomeration exists because of the substantial role of the centripetal forces.
A substantial literature (Rosenthal & Strange, 2001; De Dominicis et al., 2013; Lu & Tao, 2009) measures geographical concentration using the Ellison-Glaeser index9. This index examines the presence of localization driven by the natural advantage of specific areas and sector-specific spillovers against the localization caused by random firm-specific choices (Dauth et al., 2018). If the extent of localization is more significant than expected when firms choose their location randomly, it is concluded that the industry is geographically concentrated (Ellison & Glaeser, 1997). As spillover effects do not recognize any areal boundaries, Arbia (2001) and Lafourcade & Mion (2007) recognize simple Ellison-Glaeser (EG) index did not take into account the neighbourhood effects. Given that issue, Guimaraes et al. (2011) extend the original EG index (1997) by adding spatial dependence through a spatial weight matrix. The spatially weighted EG index is calculated below:
γ i s w = G i S − H i ( 1 − X ′ Ψ X ) ( 1 − H i ) ( 1 − X ′ Ψ X ) (1)
where G i S = ( S − X ) ′ Ψ ( S − X ) denote the spatially weighted geographical concentration ( G i ) , Ψ denotes a spatial weight matrix ( Ψ = W + I ) , where W is a weight matrix, and I represents the identity matrix10. The previous studies (Behrens & Bougna, 2015; Dauth et al., 2018; Crafts & Klein, 2021) used the spatially weighted EG index to examine the geographical concentration of industries in Canada, and Germany and the US.
To formally estimate which industries are more concentrated than what would be if firms choose their business location randomly, the following steps need to be calculated:
G = ( s j k − p j ) 2 (2)
where, s j k = x j k x k denotes the share of kth industry employment in a district (j),
while x j k represents total employment for kth industry in region j and x k
represents total employment for kth industry. Also, p j = x j x denotes the share
of total employment in a district (j), in which x j represents total employment in a spatial unit j, and x represents total employment in all spatial units (1, 2, 3, … J). Equation (2) corresponds to the geographic concentration of a particular industry in a spatial unit.
Next, we measure the expected value of the geographic concentration of a particular industry in a spatial unit when plants in an industry choose their location independently, and there are no-industry-specific natural advantages and spillovers (Ellison & Glaeser, 1997). The expected value of geographic concentration is calculated as follows:
E ( G ) = ( 1 − ∑ j x j 2 ) H , (3)
where Herfindahlindex ( H ) = ∑ k z k 2 denotes kth plant’s share of industry’s employment. Subsequently, to measure the extent of localization that is greater than what would be expected when plants choose their location randomly, calculate the difference between G and E(G) is shown below:
Excess concentration = G − E ( G ) (4)
Equation (4) helps identify districts where the industry has excess employment concentration. If excess concentration has a positive value, it indicates the extent of geographic concentration is more significant than expected to arise randomly (Ellison & Glaeser, 1997). However, when excess concentration has a negative value, the plants are more evenly distributed than expected at random.
Following the prior literature (Monmonier, 1974; Mersy, 1990), we have used the choropleth maps to evaluate the spatial distribution of workers across India’s various states and union territories. The choropleth maps depict statistical data through different colour shades on pre-defined administrative areas, i.e., states and districts. The spatial distribution includes detecting hot and cold spots, global and local patterns, comparing two or more regions’ estimated values and assessment of global and local heterogeneity. However, choropleth maps suffered from the area-size bias in which larger regions attract more attention than smaller regions (Dent, 1999; Speckmann & Verbeek, 2010). Therefore, to avoid bias, equal-area cartograms can be applied (Wood & Dykes, 2008; Kraak & Ormeling, 2020) to detect the spatial distribution of workers concerned with a particular location. Cartograms are a valuable and intuitive tool to visualize statistical data about administrative areas, where the size of an administrative area corresponds to a specific geographic variable (Olson, 1976; Van Kreveld & Speckmann, 2007; Sun & Li, 2010). Besides, Following Gastner and Newman’s (2004) diffusion-based method for constructing density-equalizing maps, we have further computed the cartogram to evaluate the spatial concentration of workers.
This study uses the latest Economic Census (EC) (2013) data, covering both formal and informal establishments across various districts in India11. It is the most acceptable micro-unit level establishment data available in the Indian context to evaluate the geographical concentration of KTI industries. Further, the EC dataset includes manufacturing and service industries’ employment of workers at the district level. The EC dataset covers 71 manufacturing and 120 service industries, followed by the National Industrial Classification (NIC-2008)12. Although the NIC (2008) provides the sub-classification of industries up to a maximum 5-digit level, EC data are available for up to 3-digit level classification. Following the EC data, we are more confined to the 3-digit level industrial classifications analysis to measure the spatial concentration of the industries at the district level data. In the EC dataset, out of 191 industries, we select 35 KTI industries, comprising 29 manufacturing and six service sector industries in India13.
Our study considers two spatial structures to measure the geographical concentration of industries. The first spatial structure measures industries’ geographical concentration using aggregated data at a district level covering 636 districts across 34 Indian states and union territories. We have to construct the spatial weight matrix to quantify geographical concentration while accounting for neighbouring effects. It requires India’s district boundaries shapefiles data, which consist of districts name, states names, geographical coordinates through latitudes and longitudes and census code14. For these 636 districts, we construct a queen contiguity weight matrix in Stata14 using the spmat command created by Drukker et al. (2013). For another spatial structure, we construct a queen contiguity weight matrix for the geographical concentration of industries within Indian states and union territories separately for every 29 contiguous states of India using district-level data. Moreover,
Our empirical interest is to evaluate the geographical concentration of KTI industries using the spatially weighted Ellison-Glaeser index (EGSPAT) at a district level in India.
Our next interest is to detect in which Indian states and union territories most of the workers of highly concentrated KTI industries are employed. As discussed, we draw a cartogram to visualize the spatial distribution of employment at the micro-unit establishment level for highly concentrated KTI industries in rural and urban India. Besides, following Jenks’s (1967) Natural Breaks optimization method, we divide the total workers employed in highly concentrated KTI industries into five levels of class intervals, depicted in five colours on a choropleth map.
| Region | Employment | Percentage (%) | Establishment | Percentage (%) |
|---|---|---|---|---|
| Rural area | 1,095,825 | 27 | 139,412 | 32 |
| Urban area | 2,902,847 | 73 | 292,431 | 68 |
| Total | 3,998,672 | 100 | 431,843 | 100 |
Source: Author’s computation using India’s Economic Census (2013) data. Notes: Economic Census (2013) provides data for 641 districts, but KTI industries have a presence in 616 and 627 districts of the rural and urban areas, respectively.
| NIC | Industry Name | EGSPAT |
|---|---|---|
| 582 | Software publishing | 0.297* |
| 309 | Manufacture of transport equipment n.e.c. | 0.252* |
| 303 | Manufacture of air and spacecraft and related machinery | 0.240* |
| 620 | Computer programming, consultancy and related activities | 0.075* |
| 302 | Manufacture of railway locomotives and rolling stock | 0.069* |
| 261 | Manufacture of electronic components | 0.060* |
| 201 | Manufacture of basic chemicals, fertilizer and nitrogen compounds, plastics and synthetic rubber in primary forms | 0.057* |
Source: Author’s computations. Notes: NIC and EGSPAT represent National Industrial Classification at a 3-digit and spatially weighted Ellison-Glaeser index. *Shows the estimated value of EGSPAT is statistically significant at a 5 percent level (Guimaraes et al., 2011). The results imply all industries are statistically significant at a 5 percent level. The EGSPAT index estimated value is greater than 0.05, indicating a highly concentrated industry.
| NIC | Industry Name | EGSPAT |
|---|---|---|
| 620 | Computer programming, consultancy and related activities | 0.571* |
| 266 | Manufacture of irradiation, electromedical and electrotherapeutic equipment | 0.387* |
| 722 | Research and experimental development in social sciences and humanities | 0.370* |
| 304 | Manufacture of military fighting vehicles | 0.218* |
| 293 | Manufacture of parts and accessories for motor vehicles | 0.204* |
| 309 | Manufacture of transport equipment n.e.c. | 0.201* |
| 262 | Manufacture of computers and peripheral equipment | 0.159* |
| 203 | Manufacture of man-made fibres | 0.090* |
| 265 | Manufacture of measuring, testing, navigating and control equipment, watches and clocks | 0.076* |
| 252 | Manufacture of weapons and ammunition | 0.070* |
Source: Author’s computations. Notes: *Shows the estimated value of EGSPAT is statistically significant at a 5 percent level. The results imply all industries are statistically significant at a 5 percent level.
workers are employed. More specifically, the choropleth map visualizes the spatial distribution of workers through the colour difference, indicating that the red colour shows a higher level of employment of workers.
In contrast,
Our subsequent empirical analysis uses district-level data to quantify KTI industries’ geographical concentration within highly concentrated states of India.
Similarly,
| NIC3 | KTI Industries | EGSPAT |
|---|---|---|
| Maharashtra | ||
| 309 | Manufacture of transport equipment n.e.c. | 0.231* |
| 201 | Manufacture of basic chemicals, fertilizer and nitrogen compounds, plastics and synthetic rubber in primary forms | 0.209* |
| 620 | Computer programming, consultancy and related activities | 0.177* |
| 302 | Manufacture of railway locomotives and rolling stock | 0.117* |
| 261 | Manufacture of electronic components | 0.056* |
| 303 | Manufacture of air and spacecraft and related machinery | 0.018 |
| 582 | Software publishing | −0.010 |
| Telangana | ||
| 303 | Manufacture of air and spacecraft and related machinery | 1.472* |
| 620 | Computer programming, consultancy and related activities | 0.277* |
| 582 | Software publishing | 0.254* |
| 302 | Manufacture of railway locomotives and rolling stock | 0.229* |
| 201 | Manufacture of basic chemicals, fertilizer and nitrogen compounds, plastics and synthetic rubber in primary forms | 0.168* |
| 261 | Manufacture of electronic components | 0.132* |
| 309 | Manufacture of transport equipment n.e.c. | −0.016 |
| Andhra Pradesh | ||
| 309 | Manufacture of transport equipment n.e.c. | 0.264* |
| 302 | Manufacture of railway locomotives and rolling stock | 0.147* |
| 201 | Manufacture of basic chemicals, fertilizer and nitrogen compounds, plastics and synthetic rubber in primary forms | 0.124 |
| 620 | Computer programming, consultancy and related activities | 0.123* |
| 261 | Manufacture of electronic components | 0.117* |
| 303 | Manufacture of air and spacecraft and related machinery | 0.083 |
| 582 | Software publishing | 0.047* |
Source: Own computations. Notes: NIC and EGSPAT represent National Industrial Classification at a 3-digit and spatially weighted Ellison-Glaeser index. *Shows the estimated value of EGSPAT is statistically significant at a 5 percent level (Guimaraes et al., 2011).
(Haryana and Tamil Nadu) in rural areas. The reported results exhibit that the manufacture of irradiation, electromedical and electrotherapeutic equipment (NIC-266) is the most highly concentrated KTI industry in Haryana. In contrast, computer programming, consultancy and related activities (NIC-620) is the top, most highly concentrated industry in Tamilnadu (estimated EGSPAT value of 0.782). As this industry belongs to the IT sector, facilitating the prevalence of the
| NIC3 | KTI Industries | EGSPAT |
|---|---|---|
| Haryana | ||
| 266 | Manufacture of irradiation, electromedical and electrotherapeutic equipment | 1.257* |
| 293 | Manufacture of parts and accessories for motor vehicles | 0.171* |
| 309 | Manufacture of transport equipment n.e.c. | 0.165 |
| 203 | Manufacture of man-made fibres | 0.113 |
| 252 | Manufacture of weapons and ammunition | 0.095* |
| 265 | Manufacture of measuring, testing, navigating and control equipment, watches and clocks | 0.070 |
| 620 | Computer programming, consultancy and related activities | 0.040* |
| 722 | Research and experimental development on social sciences and humanities | 0.009 |
| 262 | Manufacture of computers and peripheral equipment | −0.436 |
| Tamil Nadu | ||
| 620 | Computer programming, consultancy and related activities | 0.782* |
| 203 | Manufacture of man-made fibres | 0.253* |
| 293 | Manufacture of parts and accessories for motor vehicles | 0.138* |
| 309 | Manufacture of transport equipment n.e.c. | 0.124 |
| 266 | Manufacture of irradiation, electromedical and electrotherapeutic equipment | 0.078 |
| 722 | Research and experimental development on social sciences and humanities | 0.061 |
| 262 | Manufacture of computers and peripheral equipment | 0.046 |
| 265 | Manufacture of measuring, testing, navigating and control equipment, watches and clocks | 0.033 |
| 252 | Manufacture of weapons and ammunition | −0.005 |
| 304 | Manufacture of military fighting vehicles | −0.057 |
Source: Own computations. Notes: *Shows the estimated value of EGSPAT is statistically significant at a 5 percent level (Guimaraes et al., 2011).
information and Communication services industry in Tamil Nadu indicates a higher concentration of employment.
After finding evidence of the geographical concentration of KTI industries within highly concentrated states across the urban areas in India, the subsequent empirical interest is to evaluate the excess employment concentration of highly concentrated KTI industries19. More specifically, this study explores India’s urban concentrated and dispersed districts20.
Similarly,
Like in an urban area, our subsequent empirical interest is to evaluate and detect the location of excess employment concentration of highly concentrated KTI industries in rural areas across the highly concentrated states in India (Haryana and Tamil Nadu)24.
However, the dark green colour shows the plants in an industry are more evenly distributed than expected at random for the remaining districts of Haryana. Therefore, this validates that the localization effect is visible and significant in only two districts in Haryana25. Similarly,
areas, indicating the effect of localization. In contrast, in Tamil Nadu, the remaining 29 districts depicted in light and dark green patches are rural-dispersed, showing the meagre impact of localization26.
In a nutshell, our empirical findings for the spatial concentration of KTI industries are encapsulated as follows. First, we measure the geographical concentration of KTI industries in urban and rural areas using the spatially weighted Ellison-Glaeser index. The results exhibit that out of 35 KTI industries, only seven KTI industries are highly concentrated in the urban area, whereas ten KTI industries are in the rural area. Second, our subsequent empirical interest is to detect which Indian states and union territories employ a maximum number of workers in highly concentrated KTI industries. Results show that Andhra Pradesh and Maharashtra for urban areas and Tamil Nadu and Haryana for rural areas are the two Indian states that substantially employ maximum workers. Within these Indian states, only a few districts employed the highest number of workers in urban and rural areas. Third, we explore highly concentrated KTI industries’ which are more geographically concentrated within highly concentrated Indian states. The results reveal that for an urban area, within Maharashtra and Andhra Pradesh, the manufacture of transport equipment, n.e.c. (NIC-309) is a highly concentrated industry than other KTI industries. But for Telangana, the manufacture of air, spacecraft, and related machinery (NIC-303) is the top most highly concentrated KTI industry.
Similarly, for rural areas, within Haryana, irradiation, electromedical and electrotherapeutic equipment (NIC-266) is the most highly concentrated KTI industry. However, within Tamil Nadu, computer programming, consultancy and related activities (NIC-620) is the top most highly concentrated KTI industry. Finally, our subsequent empirical interest is to evaluate the excess employment concentration of highly concentrated KTI industries within highly concentrated states in urban and rural areas. Results show that in urban areas, only two districts (Pune and Thane) in Maharashtra, three districts (Medak, Hyderabad and Rangareddy) in Telangana, and no district in Andhra Pradesh shows the extent of localization. Therefore, this indicates centripetal forces play a viable role in the geographical concentration of highly concentrated KTI industries. Besides, for the rural area, only two districts (Ambala and Gurgaon) in Haryana and two districts (Kancheepuram and Erode) in Tamil Nadu are urban-concentrated. This indicates that the localization effect is more substantial and confined to specific locations in rural and urban India. In addition, the centripetal forces facilitate spatial concentration in particular locations of a few districts across the highly concentrated Indian states in the urban and rural areas.
This paper has explored the geographic concentration of knowledge and technology-intensive (KTI) industries at the district level covering 0.43 million establishments in rural and urban areas in India using the spatially weighted Ellison-Glaeser index, cartogram and choropleth map. Empirical results reveal that seven industries in urban and ten industries in rural areas are highly concentrated KTI industries in India. Further, empirical results indicate that urban areas of Andhra Pradesh and Maharashtra and rural areas of Haryana and Tamil Nadu are substantially enlarged on the cartogram, suggesting that the highly concentrated KTI industries are spatially confined to a particular location.
Moreover, results also reveal that localization is substantially more potent and geographically concentrated in particular locations of a few districts across the highly concentrated states (Andhra Pradesh and Maharashtra for the urban area; Haryana and Tamil Nadu for rural area) in India. Therefore, industries spatially concentrated in only a few locations in specific districts in India connote the effect of natural advantages or other economic forces that might increase profits for firms located near firms of the same or different industries. Successively, our results suggest that the push and pull-based supply chains and demand-based networks are pretty strong in such districts and have the potential to attract new players while entering and exploring a new market. Therefore, the policymakers may consider and emphasize highly agglomerated districts in India to boost productivity, output and employment, which would augment growth at both local and national levels. In sum, following Gertler’s (2018) ideas, for effective policy implementation, Indian policymakers may also consider multi-level governance, informal (traditions, values and conventions) and formal institutions (laws, rules and regulations), to become essential socially constructed elements for different economic actors and their economic activities.
We would like to thank anonymous journal reviewers for their valuable comments and insightful suggestions, which enabled us to improve the quality of this paper to a large extent. We would also like to thank the Editors and Joy Deng, Managing Editor of the journal, for giving us a chance to revise this paper.
The authors declare no conflicts of interest regarding the publication of this paper.
Agarwal, S., & Behera, S. R. (2022). Do Knowledge and Technology-Intensive Industries Spatially Concentrate in Rural and Urban Areas of India? Evidence from Economic Census Micro-Level Data. Theoretical Economics Letters, 12, 1095-1125. https://doi.org/10.4236/tel.2022.124060
| Manufacturing | Non-Manufacturing | |
|---|---|---|
| High R&D intensity industries | 303: Air and spacecraft and related machinery | 72: Scientific research and development |
| 21: Pharmaceuticals | 582: Software publishing | |
| 26: Computer, electronic and optical products | - | |
| Medium-high R&D intensity industries | 252: Weapons and ammunition | 62 - 63: IT and other information services |
| 29: Motor vehicles, trailers and semi-trailers | - | |
| 325: Medical and dental instruments | - | |
| 28: Machinery and equipment n.e.c. | - | |
| 20: Chemicals and chemical products | - | |
| 27: Electrical equipment | - | |
| 30: Railroad, military vehicles and transport n.e.c. | - |
Source: Galindo-Rueda & Verger (2016). Notes: NIC-26 comprises (NIC-261, 262, 263, 264, 265, 266, 267 and 268), NIC-72 comprises (NIC-721 and 722), NIC-29 comprises (NIC-291, 292, and 293), NIC-28 comprises (NIC-281 and 282), NIC-20 comprises (NIC-201, 202, and 203), NIC-27 comprises (NIC-271, 272, 273, 274, 275, and 279), NIC-30 comprises (NIC-302, 304, and 309) and NIC-63 comprises (NIC-631 and 639).
| States Name | Districts | Excess concentration |
|---|---|---|
| Maharashtra | Pune | 0.00807 |
| Maharashtra | Thane | 0.00100 |
| Maharashtra | Nashik | −0.00085 |
| Maharashtra | Mumbai Suburban | −0.00135 |
| Maharashtra | Mumbai | −0.00255 |
| Maharashtra | Nagpur | −0.00274 |
| Maharashtra | Raigarh | −0.00274 |
| Maharashtra | Kolhapur | −0.00275 |
| Maharashtra | Solapur | −0.00275 |
| Maharashtra | Amravati | −0.00276 |
| Maharashtra | Ahmadnagar | −0.00276 |
| Maharashtra | Satara | −0.00276 |
| Maharashtra | Jalgaon | −0.00276 |
| Maharashtra | Akola | −0.00276 |
| Maharashtra | Sangli | −0.00276 |
| Maharashtra | Bid | −0.00276 |
| Maharashtra | Aurangabad | −0.00276 |
| Maharashtra | Latur | −0.00276 |
| Maharashtra | Nanded | −0.00276 |
| Maharashtra | Dhule | −0.00276 |
| Maharashtra | Wardha | −0.00276 |
| Maharashtra | Parbhani | −0.00276 |
| Maharashtra | Ratnagiri | −0.00276 |
| Maharashtra | Nandurbar | −0.00276 |
| Maharashtra | Jalna | −0.00276 |
| Maharashtra | Gondiya | −0.00276 |
| Maharashtra | Buldana | −0.00276 |
| Maharashtra | Chandrapur | −0.00276 |
| Maharashtra | Yavatmal | −0.00316 |
| Maharashtra | Bhandara | −0.00316 |
| Maharashtra | Sindhudurg | −0.00316 |
| Maharashtra | Osmanabad | −0.00332 |
| Maharashtra | Washim | −0.00332 |
| Maharashtra | Garhchiroli | −0.00332 |
| Maharashtra | Hingoli | −0.00332 |
Source: Own computations.
| States Name | Districts | Excess concentration |
|---|---|---|
| Telangana | Medak | 0.1687 |
| Telangana | Rangareddy | 0.0435 |
| Telangana | Hyderabad | 0.0041 |
| Telangana | Nalgonda | −0.0028 |
| Telangana | Khammam | −0.0028 |
| Telangana | Warangal | −0.0028 |
| Telangana | Nizamabad | −0.0028 |
| Telangana | Adilabad | −0.0028 |
| Telangana | Mahbubnagar | −0.0028 |
| Telangana | Karimnagar | −0.0033 |
Source: Own computations.
| States Name | Districts | Excess concentration |
|---|---|---|
| Andhra Pradesh | Srikakulam | −0.0033 |
| Andhra Pradesh | Vizianagaram | −0.0033 |
| Andhra Pradesh | Prakasam | −0.0033 |
| Andhra Pradesh | Y.s.r. | −0.0032 |
| Andhra Pradesh | West Godavari | −0.0032 |
| Andhra Pradesh | Chittoor | −0.0032 |
| Andhra Pradesh | Sri Potti Sriramulu Nellore | −0.0032 |
| Andhra Pradesh | Anantapur | −0.0031 |
| Andhra Pradesh | East Godavari | −0.0031 |
| Andhra Pradesh | Guntur | −0.0031 |
| Andhra Pradesh | Kurnool | −0.0031 |
| Andhra Pradesh | Visakhapatnam | −0.0031 |
| Andhra Pradesh | Krishna | −0.0031 |
Source: Own computations.
| States Name | Districts | Excess concentration |
|---|---|---|
| Haryana | Ambala | 0.3336 |
| Haryana | Gurgaon | 0.1792 |
| Haryana | Rewari | −0.0053 |
| Haryana | Sonipat | −0.0143 |
| Haryana | Palwal | −0.0165 |
| Haryana | Mewat | −0.0165 |
| Haryana | Karnal | −0.0166 |
| Haryana | Hisar | −0.0166 |
| Haryana | Jhajjar | −0.0166 |
| Haryana | Jind | −0.0166 |
| Haryana | Kaithal | −0.0166 |
| Haryana | Kurukshetra | −0.0166 |
| Haryana | Panipat | −0.0166 |
| Haryana | Mahendragarh | −0.0166 |
| Haryana | Sirsa | −0.0166 |
| Haryana | Panchkula | −0.0166 |
| Haryana | Faridabad | −0.0166 |
| Haryana | Fatehabad | −0.0275 |
| Haryana | Rohtak | −0.0275 |
| Haryana | Bhiwani | −0.0275 |
| Haryana | Yamunanagar | −0.0275 |
Source: Own computations.
| States Name | Districts | Excess concentration |
|---|---|---|
| Tamil Nadu | Kancheepuram | 0.444 |
| Tamil Nadu | Erode | 0.053 |
| Tamil Nadu | Kanniyakumari | −0.006 |
| Tamil Nadu | Tiruppur | −0.01 |
| Tamil Nadu | Thiruvallur | −0.014 |
| Tamil Nadu | Theni | −0.015 |
| Tamil Nadu | Thoothukkudi | −0.015 |
| Tamil Nadu | Dharmapuri | −0.015 |
| Tamil Nadu | Thiruvarur | −0.015 |
| Tamil Nadu | Pudukkottai | −0.015 |
| Tamil Nadu | Nagappattinam | −0.015 |
| Tamil Nadu | Thanjavur | −0.015 |
| Tamil Nadu | Madurai | −0.015 |
| Tamil Nadu | Karur | −0.015 |
| Tamil Nadu | Tiruchirappalli | −0.015 |
| Tamil Nadu | Coimbatore | −0.015 |
| Tamil Nadu | Tiruvannamalai | −0.015 |
| Tamil Nadu | Virudunagar | −0.015 |
| Tamil Nadu | Vellore | −0.015 |
| Tamil Nadu | Viluppuram | −0.015 |
| Tamil Nadu | Tirunelveli | −0.015 |
| Tamil Nadu | Dindigul | −0.015 |
| Tamil Nadu | Krishnagiri | −0.015 |
| Tamil Nadu | Salem | −0.015 |
| Tamil Nadu | Sivaganga | −0.015 |
| Tamil Nadu | Ariyalur | −0.017 |
| Tamil Nadu | Ramanathapuram | −0.017 |
| Tamil Nadu | Cuddalore | −0.017 |
| Tamil Nadu | Namakkal | −0.017 |
| Tamil Nadu | Perambalur | −0.215 |
| Tamil Nadu | The Nilgiris | −0.217 |
Source: Own computations.