Abstract
This article empirically examines the existence of inter-sectoral growth linkages among the key sectors of the Indian economy at the state level. The examination evaluates the impact of the non-agricultural sectors of the states and that of the rest of the states on agricultural output of a particular state. An annual panel data set for 15 general category states have been taken for the period 1980–1981 to 2012–2013. Panel cointegration and fully modified ordinary least square methods have been used to study the existence of a long-run equilibrium relationship between sectors. The results suggest that there is a long-run equilibrium relationship among three sectors of the economy in the Indian states. The evaluation indicates that the industrial sector contributes positively in complementing the growth of agriculture, but the service sector advancement affects agricultural growth negatively. However, services having some direct reference to agriculture such as transport, storage and communication (TSC), trade, hotel and restaurant (THR) and banking and insurance (BI) have positive linkage with agriculture. The state specific econometric evaluation of the agricultural output varies relatively across different states, for example, in Kerala, the impact of rest of the industries and services leaves a positive significance; whereas, the study foresees the negative impact of industry and services in the states such as Bihar, Madhya Pradesh, Orissa and Rajasthan. In order to neutralize the negative linkages of service sector on agriculture, policies for promoting pro-agricultural services such as crop and agricultural insurance, agricultural loans, facilities for agricultural warehouse, marketing services, weather communication, transport services and provision of technical support to farm activities are important. Such initiatives can help agricultural sector grow along in the simultaneous development of sectors propelling growth of the economy at a faster rate.
Introduction
Inefficiency in the sectoral reallocation of resources across nations going through different stages of growth trajectories has demanded considerable research attention on the process of structural transformation. Structural transformation, which refers to the broad reallocation of resources across the main sectors such as agriculture, industry and services is considered an important aspect of the process of economic growth (Herrendorf, Rogerson, & Valentinyi, 2014). With the advancement of economy, the sectoral contribution to gross domestic product (GDP) moves in favour of industry and services. Although seminal contributions to this topic have been made by pioneers, namely, Clark (1940), Chenery (1960), Kuznets (1973) and Syrquin (1988), there still remains enough room to explore the specific traits exhibited through this phenomenon in the context of a developing country like India.
Emerging economies have registered a delayed entry to catch up to the wagon of modern economic growth triggered by industrial revolution and economic reforms. As different from the case of the developed nations which have already gone through the initial phases of structural transformation, emerging economies of the postcolonial era have experienced considerable changes in their consumption pattern, demand pattern, composition of sectors and social indicators. India is no exception to this pattern of changes that is witnessed in the process of structural transformation.
In the Indian context, it was the initiation of the reforms in the mid-1980s and its implementation in the 1990s which was responsible for increasing the share of services in GDP. Figure A.1 shows the sectoral GDP of agriculture, industry and services sector. The agricultural contribution has been declining over the years; whereas, the share of service sector in GDP has an upward trend. Figure A.2 shows the decadal average of sectoral contribution of agriculture to gross state domestic product (GSDP) for Indian states which has also declined over the last three decades. There is a structural transformation in India in favour of service sector.
However, the importance of agricultural sector in India cannot be overlooked as more than 50 per cent of the population still depends on it. Agriculture is fully integrated with the other sectors of the economy. It not only provides raw materials to industry but also helps in creating surplus that can be used for capitalizing other sectors of the economy. It also helps in achieving self-sufficiency in food grains. Growth of agricultural sector is also important for the industrial sector. Industrial and Agrarian revolutions always go together, and economies in which agriculture is stagnant do not show industrial development (Lewis, 1954). The real challenge for an emerging economy like India is the growth of agriculture and industry along with the services sector and sustaining them. If sectors are linked to each other, then a simultaneous advancement of all sectors is possible and faster economic growth can be attained. In the development process of the Indian economy, taking the agricultural sector along with the advancement of other sectors is important
Therefore, in order to accelerate economic growth, it becomes important to verify how the non-agricultural sectors influence agricultural sector and identify the strategies to promote linkages among them. Accordingly, the main objective of the article is to examine the inter-sectoral linkages among the three sectors such as agriculture, industry and services in India’s economy and their implication on agriculture. The study verifies whether the development of non-agricultural sector of a particular state and of the rest of the states contributes to the development of agriculture in a state. The results confirm the existence of a long-run equilibrium relationship across different sectors of the economy. It is observed that the industrial sector contributes positively in complementing to the growth of agriculture, but the service sectors’ advancement affects agricultural growth negatively.
The remainder of the article is structured as follows. The review of literature pertaining to the subject is presented in Section II; Section III describes data and methodology; Section IV deals with the empirical results and discussion and conclusion of the study is given in Section V.
An Overview of Literature
The concept of sectoral linkages has evolved from Hirschman’s unbalanced growth theory. Most of the advanced economies clearly show that development has followed a series of unbalanced growth, with growth being transferred from the leading sectors of the economy to its followers. Hirschman therefore proposes a large-scale investment in certain strategic areas of development. The initial dis- equilibria faced will induce investment in other areas. The unbalanced growth theory states that those sectors with the highest linkages will be the ones that will be responsible for the generation of growth, revenue and employment (Hirschman, 1958).
The linkages among the sectors are generally examined using the Leontief input–output model and econometric modelling. Linkages among sectors of an economy have been studied by Dhawan and Saxena (1992), Sastry, Singh, Bhattacharya, and Unnikrishnan (2003), Hansda (2006), and Kaur, Bordoloi, and Rajesh (2009) using a Leontief input output–model. Dhawan and Saxena (1992) have used a supply side input–output model, along with the demand side, to study the interdependence across industries. Accordingly, industries have been classified into four categories. The first group consists of chemicals, metal products, machinery, petroleum and coal products which are those industries that have very high forward and backward linkages. These sectors contributed to roughly around 21.7 per cent in 1973–1974; whereas in 1978–1979 and 1983–1984, the figure has increased to around 28.3 per cent. Second in the priority with high forward and low backward linkages come machinery, rubber, plastic, transport goods, construction and services. In the year 1973–1974, this sector contributed to around 39.1 per cent of the total and in the years 1978–1979 and 1983–1984 declined to 30.4 per cent. The third set comprises those sectors that have low forward and high backward linkages. In the year 1973–1974, this set comprised coal and lignite, crude petroleum, natural gas and minerals excluding iron ore. In addition to the three sectors, banking and insurance and wood-based industries were added. The fourth category consisting of crops, food products, services, fishing, forestry, logging and non-metallic mineral resources sectors contributed to around 30 per cent of the total. These sectors had low forward and low backward linkages.
Sastry et al. (2003) have used the input–output model and a simultaneous linear framework for studying the sectoral linkages. They have observed that in spite of the growing tertiarization of the Indian economy, agriculture contribution to other sectors is inevitable. The existence of production linkages was seen in the 1960s; whereas in the 1990s, there existed demand side linkages across the sectors. Hansda (2006) have used Rasmussen’s indices to study inter-sectoral linkages. The sectors that contributed the highest in the total (backward and forward linkages) are trade, other transport and services, construction and other crops. Kaur et al. (2009) have observed that the demand for the inputs from industry is more for the services rather than from the agricultural sector when analysing the linkages from the production side. Whereas, the demand side linkages show that the agriculture sector has a greater degree of association with the industry.
Some studies have used econometric models to estimate the extent of inter-sectoral linkages. Feder (1983) had defined a model where output of a sector is a function of labour, capital and the externalities from the other sectors. Gemmell, Lloyd, and Mathew (2000) have used the Feder model in a three-sector framework to analyse the inter-sectoral agricultural linkages. For the Malaysian economy, manufacturing had a positive impact on agriculture, agriculture had a negligible impact on the other two sectors, and services on the other hand had a negative impact on the agriculture. Bhattacharya (2004) has analysed the sectoral growth of agricultural output across Andhra Pradesh, Karnataka and Uttar Pradesh. The variable rainfall has a positive impact on agricultural growth for the three states. Irrigation and fertilizer are positive and significant in affecting agriculture for the state of Uttar Pradesh. Industrial sector output on the other hand has been positively affected by the lagged values of agricultural output for the three states. Public expenditure and service sector growth have positive impacts on agricultural production for the state of Karnataka. Industrial growth has been taken as a variable for the states of Andhra Pradesh and Uttar Pradesh and found to be positively significant in contributing to the service sector growth. Service sector output has been modelled with the commodity sector output, growth of services and industrial growth. The service sector growth has been positively significant for the states of Andhra Pradesh and Karnataka. Industry has been significantly contributing to the service sector for all the three states.
Kaur et al. (2009) have observed a long-run association among the main sectors. At the sub-sectoral level, the long-run cointegration relationship has been observed between banking and insurance, manufacturing and the primary sector. Trade, hotel, transport and communication also exhibited a long-run association. Behera (2012) have used Granger causality and cointegration to study the linkages among industry, agriculture and service sectors for the state of Odisha. A bidirectional causality has been observed between the primary sector and the trade, hotel and restaurant services. The secondary sector has a unidirectional causality with finances, real estate and business services. A long-run relationship has been observed between the secondary sector and transport, storage and communication. The sub-sectors such as finance, real estate and business services were found to be cointegrated.
Studies of Bhide, Chadha, and Kalirajan (2006) and Chakravarthy (2006) have looked at the concept of linkages across states in an open economy framework. The open economy framework considers the interaction across states. It verifies the existence of spillovers across states that is how the growth in a state is affected by the growth of other states. Bhide et al. (2006) have used Granger causality to study the causation among the states. In their study, Kerala appears to be the most influenced state as growth in the states of Rajasthan, West Bengal and Odisha has influenced the Kerala’s growth rate. The second group of states are the states that have been influenced by the growth rate of two other states. The states include West Bengal, Haryana, Karnataka, Andhra Pradesh and Rajasthan. The next group of states Assam, Bihar, Punjab, Tamil Nadu and Madhya Pradesh have been influenced by one state. Lastly Maharashtra, Odisha and Uttar Pradesh have not been influenced by the growth rate of any other state.
Chakravarthy (2006) has done a demand side analysis using the log-linear model for the service sector growth of states. The service sector growth rate is a function of non-service sectors of the state and rest of the states. Industry emerged to be an important determinant affecting agricultural growth. Prior to the reforms the non-service sectors of other states have been significant for the states Haryana, Gujarat, Madhya Pradesh, Uttar Pradesh, Bihar and Kerala. Whereas, in the reform period, it has been significant for the states of Haryana, Maharashtra, Rajasthan and Madhya Pradesh.
From the literature it is observed that though there are a number of studies that have analysed the inter-sectoral linkages across sectors, sparse studies address the issue of sectoral linkages across states in an open economy framework. Besides, the present article looks at the implication of sectoral linkages for agricultural sector. It is important to study inter-sectoral linkages for the agricultural sector as it will help in capturing the growth impulses that the sector had experienced due to the structural transformation of the economy.
Data and Methodology
Data
Secondary data have been taken from Central Statistical Organization (CSO) for the 15 general category states 1 for the period 1980–1981 to 2013–2014. The special category states have not been included considering that their funding resource pattern is different from general category states. The variables include agricultural gross state domestic product (AGSDP), industrial gross state domestic product (IGSDP), service sector gross state domestic product (SGSDP), and the rest of the states’ industrial and service sector gross state domestic product (ROSISGSDP). The ROSISGSDP is the aggregate of IGSDP and SGSDP of all the states excluding the state under study. For finding out the key sub-sectors (services) that contribute to the advancement of agriculture, we have used data relating to transport, storage and communication (TSC), trade, hotel and restaurants (THR) and banking and insurance (BI). The GSDP data for all components of previous series has been spliced in order to harmonize with the 2004–2005 base year. All variables have been changed to the 2004–2005 base year at constant prices. Logarithmic transformation has been performed in order to reduce the skewness in the distribution.
Methodology
In order to verify the time series properties, panel unit root test developed by Im, Pesaran and Shin has been used for the purpose. Findings in Table 1 shows that the variables are stationary at first difference.
Panel Unit Root Test (Im Pesaran and Shin)
Panel Unit Root Test (Im Pesaran and Shin)
Pedroni panel cointegration test is employed to verify the long-run equilibrium relationship between the variables (Pedroni, 2004). The short-run dynamics has been established using panel causality test developed by Dumitrescu and Hurlin (2012). Finally, fully modified ordinary least squares (FMOLS) and dynamic ordinary least squares (DOLS) have been used in order to estimate long-run coefficients for verifying the inter-sectoral linkages. The FMOLS helps capture the dynamics and correct for serial correlation (Phillips & Hansen, 1990). A simple model has been used where agricultural output of a state is a function of industry, services of that state and the rest of the states’ industry and services. This helps verify the impact of non-agricultural sectors on agricultural output. The model is given as in Equation (1).
Sectoral Linkages in a Panel Framework
This section provides empirical results and discussion pertaining to analysis of long-run relationship among the key sectors of the economy. The Pedroni panel cointegration test has been used to test the existence of long-run sectoral linkages among agriculture, industry and services in Indian states.
Table 2 shows the results of the Pedroni panel cointegration test. Majority of the test statistics reject the null hypothesis of no cointegration among the sectors. Therefore, even if there are any short-run discrepancies among the variables in the long run, they eventually converge. In the long run, results confirm the existence of inter-sectoral linkages among the main sectors of the economy at sub-national level. For checking the robustness other cointegration tests such as Fisher and Kao have been performed which also confirms the existence of a long-run equilibrium relationship.
Pedroni Panel Cointegration Test
Pedroni Panel Cointegration Test
Table 3 shows the direction of the causality. Findings show the existence of bidirectional causality between IGSDP and AGSDP. Whereas a unidirectional causality existed between SGSDP and AGSDP in the short run. In the short run, services cause AGSDP. The variables AGSDP and ROSISGSDP showed the existence of independence.
Panel Causality
The long-run output elasticity for agriculture has been calculated using FMOLS and DOLS. These techniques will provide relatively efficient estimate in the presence of serial correlation. Table 4 shows results of FMOLS and DOLS.
Long-run Coefficients of Agricultural Output (AGSDP)
Findings indicate that the industrial sector has a positive and significant relationship; whereas the service sector has a negative and significant relationship with agriculture. For 1 per cent change in industrial output, agricultural output increases by 0.39 per cent. The sign of coefficients remains the same even for an alternate estimation technique like DOLS. Industrial sector emerged to be a major determinant of agricultural output as there is a rapid demand for industry-based commodities for the development of agriculture (Kaur et al., 2009). It provides inputs for agricultural growth such as advanced tools, electricity, fertilizers and chemicals. Industry has also helped in increasing the demand for wage goods (Satyasai & Viswanathan, 1999). On the supply side, agriculture provides raw materials and output for agro-based industries. There is a strong linkage between agriculture and industry. However, the service sector linkages for agriculture are also important as it contributes to more than 60 per cent of GDP. In the present study service sector emerged to be significant, but the coefficient is found to be negative in influencing agricultural output. For 1 per cent change in services, agricultural output decreases by −0.01 for both FMOLS and DOLS estimations. The contribution of agriculture in the total GDP has been decreasing due to a lesser elasticity of service sector output to agriculture. The main reasons for the same might be that the service components such as agricultural loan, insurance, transport, hotels and restaurants are not properly emphasized. The rest of the state’s non-agricultural output has not been significant for agriculture in the panel data.
State Specific Modelling of Agricultural Growth
Sectoral Linkages of Specific Services with Agriculture
As the overall service sector provides negative coefficient for agriculture, an attempt has been made to verify the cointegration of some of the key service sectors having implication with the agriculture sector. The three sub-sectors considered are TSC, THR and BI. These services have relevance to agriculture, either they provide service inputs to agriculture sector or agricultural products are used by them. Pedroni panel cointegration test has been used to study the existence of a long-run relationship.
The result of which has been reported in the appendix. Table A1 shows the existence of a long-run relationship between the TSC and agriculture. Table A2 shows the existence of a long-run relationship among the variables, agricultural growth and BI. A long-run relationship between THR and agricultural growth rate has been established as shown in Table A3. The results suggest that if we focus on services which can help promote agriculture then the service sector growth can take along with agriculture.
Since the cointegrating relationship has been established, we move on to the panel causality test. Table A4 shows that the direction of the causality in the short run, runs from transport to the service sector. There exists a unidirectional causality from the banking BI to the agriculture sector as shown in the Table A5. There exist bidirectional causalities across the variables, THR and agriculture as shown in the Table A6.
In order to quantify the relationship, we have used the fully modified ordinary least squares. Table A7 shows the relationship between agriculture and TSC; Table A8 shows the relationship between agriculture and BI. Third, Table A9 shows the relationship between agriculture and THR. The coefficients clearly show a positively significant relationship for all the three models used. The THR has a relatively higher coefficient than the other two components of services. So, if services in these sectors are promoted, then the agricultural output will witness a higher growth.
State Specific Analysis
In order to understand the co-movement of principal sectors in the economy at state level, separate regressions have been estimated for each state using time series data. Results are reported in Table 5. This is very important as it helps in studying the impact of IGSDP, SGSDP and ROSISGSDP on the AGSDP of a particular state.
It reflects that the industrial growth has a positive and significant effect on agriculture for the states such as Goa, Haryana, Punjab, Kerala, Bihar and Uttar Pradesh. However, industrial output has a negative impact on agriculture for the states such as West Bengal and Odisha. The service sector has a negative influence for the states such as Punjab, Kerala and Bihar; whereas, it has shown positive significance for the states such as Andhra Pradesh, West Bengal, Uttar Pradesh, Odisha and Madhya Pradesh. One important factor for the negative coefficient of these states is the potential labour migration both outward and to non-primary sectors. The rest of the states’ non-agricultural sector contribution has significantly and positively affected the growth of agriculture for the states like Kerala. The results reflect a negative significance in the states such as Bihar, Madhya Pradesh, Odisha and Rajasthan. One pattern is evident that with the advancement of non-agricultural sectors in the rest of the states, the agricultural sector in Kerala is growing while that of low-income states is witnessing a negative pattern. This is mainly because of the migration of agricultural labourers from low-income states to agricultural sector in Kerala and to non-primary sectors of the rest of the states. Ghosh (2010) observed that Uttar Pradesh and Bihar are the largest sending economies followed by Odisha and Rajasthan. The prominent reason as given by the World Bank is that these states are a part of the low-income category of states that continue to lag behind the rest of the country (The World Bank, 2018). Apart from this, a few more factors which have adversely affected the agricultural growth in low-income states are brain drain, lack of adequate capital investment in the sector and lack of industrialization. A revision is also required in the fields such as agrarian structure, governance, physical and economic infrastructure and industrial policy (Rasul & Sharma, 2014). Another reason that has affected negatively in the growth of agricultural production is the mobility of labour in inter- and intra-sectors and the gaining importance of employing labour on a contract basis (Auer & Jha, 2009; Sharma, 2006).
From the state level analysis, it is observed that the service sector output has positive influences for some states such as Andhra Pradesh, West Bengal, Uttar Pradesh, Odisha and Madhya Pradesh, while it is negative for three states such as Punjab, Kerala and Bihar. But the panel data results for the overall analysis in section ‘Sectoral Linkages in a Panel Framework’ suggest a negative linkage between service sector output and agricultural output of states. Estimates from the panel data models are more robust as the models control for group-specific and time-specific heterogeneity, and degrees of freedom are much higher than the same in a single dimensional cross-sectional or time-series estimates. Sometimes, on time-series data for individual states may not provide a particular pattern, but panel data may give. Although, the impact of services sector is positive for agriculture for some states, considering the precision of the panel data estimates, the inferences for states in general is based on the findings drawn from the panel data.
The main objective of the article was to examine the inter-sectoral linkages and verify the influence of non-agricultural sectors on agricultural sector. Findings suggest the existence of a long-run equilibrium relationship among the key sectors of the economy such as agriculture, industry and services. The evaluation indicates that the industrial sector contributes positively in complementing to the growth of agriculture, but the service sectors advancement affects agricultural growth negatively. However, taking the components of services which are directly linked to agriculture such as BI, TSC and THR have a significant positive influence on agriculture.
The non-agricultural sectors of other states do not show a significant relationship. Whereas, the state level analysis shows that the rest of the states’ non- agricultural sector variable has been positively significant for Kerala and negatively significant for low-income states of Bihar, Rajasthan, Odisha and Madhya Pradesh. With advancement of non-agricultural sectors in rest of the states, there is a decline in output of agriculture in low income states. Labour retaining strategy and technological adoption in agriculture in these states will promote agriculture.
The output elasticity coefficient shows that Industry sector growth has emerged as an important determinant of agricultural growth across states. So, for the development of agriculture sector in the long run, industrial sector is important. In low-income states such as West Bengal and Odisha industrial sector has adversely affected agriculture. Agro-based, small-scale and cottage industries in these states need to be promoted to link to positive growth of agriculture. Second, promoting agriculture with pro-agricultural services such as crop and agricultural insurance, agricultural loans, facilities for services of agricultural warehouse, weather communication, transport, hotel and restaurants will help in the simultaneous advancement of agricultural sector along with the growth of service sector.
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.
Footnotes
Appendix
Acknowledgements
The authors are grateful to L Venkatachalam, Sigamani P and N Rajagopal for their comments that greatly improved the manuscript. Thanks are due to the faculty members in the Department of Economics, Central University of Tamil Nadu for their comments and suggestions during the review seminar. The first author is grateful to University Grants Comission for providing with the fellowship to pursue PhD.