Labour Market Adjustment and Intra-Industry Trade: Empirical Results from Indian Manufacturing Sectors

Abstract

During the last two decades, India has witnessed several trade and industrial policy reforms. The objective of the study is to examine the relationship between dynamism of India’s two-way trade, measured through Marginal Intra-Industry Trade (MIIT) index, and labour market adjustments, reflected through absolute employment changes, in select manufacturing sectors over 2001–2015. India’s MIIT in select sectors generally display an upward trend over the sample period, while a mixed dynamics is observed on the employment front. The generalized method of moments (GMM) estimation results indicate that MIIT, increase in productivity, skilled workforce intensity, industrial concentration, incremental FDI inflows and trade openness positively influence absolute employment changes, whereas unskilled wage exerts a negative impact on the same. The analysis further concludes that high relative growth rate, skill-intensity, incremental FDI inflows and higher productivity in a sector, also characterized by higher MIIT, may lead the firms to employ more productive and competitive resources, resulting in higher absolute changes in employment. The obtained results do not support the Smooth Adjustment Hypothesis (SAH) predictions in the Indian context.

Keywords

Marginal Intra-Industry Trade trade and labour market interactions India empirical estimation GMM Smooth Adjustment Hypothesis

Introduction

The classical theories of Smith and Ricardo as well as the Heckscher–Ohlin theorem noted ‘specialization’ as an outcome of trade liberalization, indicating possible adjustments and reallocation of resources across sectors within trading partners. However, incomplete specialization (Walde, 1994) and two-way trade within sectors dominate the real-world scenario. The expression ‘intra-industry trade’ (IIT) was initially used by Balassa (1966) for describing the phenomena of simultaneous export and import of products within an industrial classification (e.g., SITC three-digit level). Early empirical works on this branch of literature focused both on the measurement issues and pattern of IIT (Dreze, 1960, 1961; Grubel, 1967; Grubel & Lloyd, 1971, 1975; Verdoon, 1960). The subsequent focus moved towards the decomposition of the IIT in sub-categories, namely, ‘horizontal’ (HIIT) and ‘vertical’ (VIIT), and analysing their determinants. While HIIT results from the demand for product varieties, VIIT may emerge from quality differentiation (Aturupane et al., 1999; Greenaway et al., 1995; Abd-el-Rahman, 1991; Zhang et al., 2005).

Since the beginning of General Agreement on Tariffs and Trade (GATT) negotiations, multilateral, regional as well as unilateral reform processes have led to significant reduction of tariff barriers across countries (Bown & Irwin, 2016). Trade liberalization is likely to increase the share of IIT in a country’s trade flows (Balassa, 1966; Globerman & Dean, 1990). Liberalization measures may increase export potential through two routes. On one hand, intra-firm orientation may shift from inefficient to efficient activities (Bacchetta & Dellas, 1997). On the other hand, enhanced global competition may lead domestic industries to specialize in manufacturing of unique varieties to exploit scale advantages and harness efficiency gains (Melitz, 2003). At the same time, import of quality raw materials may increase, following reduction of tariff and other barriers to imports. Such efficiency gains, coupled with possible rise in imports in post-reform period, may jointly contribute to rising IIT levels.

The reforms might, however, be associated with adjustment costs, which depend on whether prevailing trade is of an inter-industry or intra-industry type in nature. The relationship between IIT and industrial adjustment is rich with both empirical and theoretical perspectives. Balassa (1966) underlined that industries with high levels of IIT undergo less structural changes in the post-liberalization period as compared to industries characterized by low levels of IIT. Krugman (1981) demonstrated that when similar factor endowments prevail in trade partners, both countries tend to gain from liberalization, enhanced two-way trade and the consequent rise in IIT, with fewer structural adjustment problems. Employment and production disruption is therefore minimized when adjustment process is internal to an industry as it is easier to transfer and adapt resources (e.g., labour) among firms within the same product group (Balassa, 1966). As a result, lower adjustment cost is involved when trade is of IIT-type as compared to inter-industry type (Brulhart & Elliott, 2002; Brulhart et al., 2006; Greenaway & Milner, 1986). Conversely, trade liberalization leads to greater structural adjustment in industries characterized by low levels of IIT (Hamilton & Kniest, 1991). The literature on interrelationship between IIT and labour market adjustment effects is known as the ‘Smooth Adjustment Hypothesis’ (SAH) and a dynamic measure of IIT, namely the Marginal IIT (MIIT) index, is commonly used for this purpose (Brulhart, 1994; Brulhart et al., 2006; Greenaway et al., 1994; Lovely & Nelson, 2000, 2002). The adjustment effect has been found to be weaker for the VIIT-type rather than HIIT-type industries (Brulhart & Elliott, 2002; Devadason, 2012).

Increased IIT and associated labour market adjustments can be explained through the specific factors open economy framework developed by Neary (1985). As enhanced trade flows through liberalization reduce the relative price of importable goods, the factor market would undergo a series of adjustment costs, if the process is not perfectly smooth. First, unskilled workers can try to move between sectors but as the wage rate is sticky in downward direction, excess supply of workers may lead to temporary unemployment. Second, while full employment is automatically ensured in presence of perfectly flexible wages, the workers may still temporarily remain tied to their sectors, given the adjustment considerations (e.g., job-searching, retraining and geographical location costs). Hence, the market for skilled labour can become segmented in the short run with temporary inter-sectoral wage divergences (Baldwin et al., 1980).

Reform measures were introduced in India in mid-1980s, though significant changes followed only after the launch of the liberalization programmes in 1991 (Panagariya, 2004). Policy reforms led to manufacturing productivity growth in the country (Deb & Ray, 2014), with positive implications for exports. On the other hand, the backward and forward integration of Indian manufacturing sectors has deepened over the years (Nag, 2016). The growing trade and investment integration with the world has led to rising trade overlap and IIT levels for the country (Aggarwal & Chakraborty, 2017; Burange et al., 2017). It has been observed that the aggregate IIT index has increased for India from 42.08 to 57.58 over 2001 to 2018 (Table 1). In addition, the sectoral IITs for key manufacturing segments have also witnessed an upward movement (Aggarwal & Chakraborty, 2019). The increase in IIT index has been caused by the growing development level in the country (Banerjee & Bhattacharyya, 2004). Several Indian manufacturing sectors have been characterized by the phenomenon of ‘jobless growth’ even in the post-reform period (Datt, 1994; Mazumdar & Sarkar, 2004; Mohan, 2014). On the other hand, the country has witnessed a widening trade deficit in several manufacturing sectors, with consequent output and employment repercussions (Chaudhuri, 2015). The growing manufacturing trade deficit and the associated domestic compulsions had been one of the major reasons behind India’s recent pull-out from the Regional Comprehensive Economic Partnership (RCEP) negotiations (Dhar, 2019) and the announcement of the ‘Atmanirbhar Bharat Abhiyan’ in May 2020 (GoI, 2020). There is a need to understand how the dynamics in sectoral IIT patterns might influence the corresponding labour market adjustment processes, as the interrelationship between the two in Indian context remains relatively unexplored in existing literature.

Table 1.

Marginal IIT in India’s Trade for Select Product Groups and Years

Commodity Sections	HS Codes	GL Index					B Index
Commodity Sections	HS Codes	2001	2005	2010	2015	2018	2001	2005	2010	2015	2018
Chemicals	28, 29	45.81	54.07	55.17	63.52	72.34	22.98	62.04	90.09	50.68	75.37
Leather and footwear	41, 42, 64	29.21	38.51	38.38	53.49	64.56	10.54	48.45	71.21	8.12	65.87
Textile and garments	50–63	18.29	26.35	32.99	38.65	45.98	22.76	38.22	18.08	17.77	18.52
Iron and steel	72, 73	49.73	48.91	44.00	50.72	56.62	8.44	28.61	23.23	20.86	46.38
Base metals	74–83	34.04	41.21	33.99	44.15	49.34	38.96	92.23	64.89	52.45	48.45
Electrical and electronics machinery and equipment	84, 85	65.54	57.60	67.07	54.95	58.90	26.33	42.33	24.35	18.35	45.94
Vehicles and auto-components	87	51.98	55.10	40.77	41.55	55.38	29.19	45.20	47.07	24.11	56.59
Aggregate		42.08	45.96	44.62	49.57	57.58	36.78	73.71	68.58	87.30	59.97

Source: The authors (computation from Trade Map data, International Trade Centre [n.d.]).

Given this background, the present article intends to analyse the influence of IIT on employment changes in seven select manufacturing segments in India over 2001–2015. The analysis is arranged along the following lines. First, a brief review of MIIT literature is presented, followed by the evidence on IIT in India. The empirical model and data for the analysis are outlined next. Finally, based on obtained results, certain policy conclusions are drawn.

Measurement of Marginal Intra-Industry Trade

The literature on measuring MIIT is quite rich. The first set of analyses involves a volume-based approach, where a dynamic MIIT index is constructed to investigate the relationship between trade changes and the cost of adjustment associated with it. The approach initially involved comparison of Grubel-Lloyd (GL) indices for different time periods. Hamilton and Kniest (1991) criticized this practice on two counts. First, an increase in inter-industry trade flows may result in an increase in IIT (GL index), when such increase reduces the trade imbalance in the sector under consideration. Second, the inter-temporal comparison of GL indices does not facilitate conclusions on the structure of the change in trade flows but merely indicates changes in the trading pattern. In light of the aforesaid limitations, Hamilton and Kniest (1991) proposed the following measure of MIIT, commonly known as MIIT_HK:

M I I T_{H K} = {\begin{matrix} \frac{X_{t} - X_{t - n}}{M_{t} - M_{t - n}} ​ ​ ​ f o r M_{t} - M_{t - n} > X_{t} - X t - n_{t} > 0 \\ \frac{M_{t} - M_{t - n}}{X_{t} - X_{t - n}} ​ ​ ​ f o r X_{t} - X_{t - n} > M_{t} - M t - n_{t} > 0 \\ u n d e f i n e d ​ f o r X_{t} \leq X_{t - n} o r M_{t} \leq M t - n_{t}, \end{matrix}

where X_t/M_t and X_t_-n/M_t_-n are exports and imports of a particular industry in period t and t−n, respectively, and n stands for the number of years. This measure examines the structure of the change in trade flows and therefore, eliminates the shortcoming of the earlier method.

The analysis of Hamilton and Kniest (1991), involving all manufacturing sectors at four-digit level of the Australian Standard Industrial Classification (ASIC), tests the relationship between IIT and trade liberalization as well as the associated structural adjustment. The result provides evidence that trade liberalization leads to lower structural adjustment in industries with high levels of IIT. However, Greenaway et al. (1994) noted that the HK index can result in a non-random omission of a significant number of statistical observations and therefore, to potentially misleading interpretations, as the index is undefined when either X or M has decreased in comparison to past period. Analysing the SITC three-digit industry-level data for UK, Greenaway et al. (1994) proposed MIIT indices for more effectively evaluating the adjustment implications of trade expansion, known as MIIT_GHME:

M I I T_{G H M E} = {[(X + M) - | X - M |]}_{t} - {[(X + M) - | X - M |]}_{t - n}

M I I T_{G H M E} = Δ [(X + M) - | X - M |]

where X and M stand for exports and imports of a particular industry during years t and t−n, and n stands for the number of years.

However, it was noted that Hamilton and Kniest (1991) criticism of the GL-comparison method also applies to the GHME measure, since the latter is always defined but closely resembles the GL indices and fails to capture the structure of change in trade patterns (Brulhart, 1994). Two major criticisms have been levelled against the GHME measure. First, the GHME measure reports IIT in absolute values rather than as a ratio unlike the GL and HK indices. As an unscaled number, it lacks presentational appeal of standard indices, which lies between 0 and 1. Second, it is silent on the proportion of MIIT relative to inter-industry trade, thereby providing an inferior result (Brulhart, 1994).

An alternative measure developed by Brulhart (1994) reveals the structure of the change in import and export flows, like the HK index. Moreover, it is defined for all cases and shares all the familiar statistical properties of the GL index. The dynamic MIIT index propounded by Brulhart (B) is calculated as:

B = \frac{{\sum^{​}}^{​} [(| Δ X_{i t} | + | Δ M_{i t} |) - | Δ X_{i t} - Δ M_{i t} |]}{{\sum^{​}}^{​} (| Δ X_{i t} | + | Δ M_{i t} |)} \times 100

(1)

This index varies between 0 and 100 where 0 and 100 represent marginal trade in the particular industry to be completely of inter-industry and IIT type, respectively. The index possesses several advantages (Brulhart, 1999; Brulhart & Elliott, 1998). First, it can be observed that a country where exports and imports in a particular sector grow or shrink at a similar pace (high B), intra-industry geographical specialization is likely to occur while the overall performance of the sector is determined by global demand or technology changes. On the other hand, inter-industry geographical specialization is likely to occur in case a country’s exports and imports in a particular period show a diverging trend (low B). Second, sectoral performance is defined as the change in exports and imports in relation to each other, where they represent strengthening and weakening of the particular industry, respectively. So, the magnitude of change in exports and imports determine the value of B. Given these advantages, a number of studies have used the B index to explore the relationship between IIT and industrial adjustment (Brulhart & Thorpe, 2000; Veeramani, 2004).

Azhar and Elliott (2004) have proposed another index (S) for measuring adjustment effects, which combines the two strands of literature; namely, volume-based and quality-based approach to measure the changing structure of product quality associated with changes in IIT. The current analysis, however, measures MIIT using the B index.

One important methodological issue deserves mention here. Measurement of MIIT indices for empirical analysis requires a choice of the most appropriate time period as relative timing of trade affects the calculations (Brulhart, 2000). Oliveras and Terra (1997) investigated the statistical properties of the B index and pointed out that there is no systematic relationship between the B index calculated over a certain time interval and corresponding indices over constituent sub-intervals. Similar absence of interrelationship was also observed between the B index of a given industry and the corresponding indices of the sub-industries. The systematic link between the overall B index and its sub-interval components exists only if there is a continuous improvement/deterioration of the trade balance for the aggregate time interval and its sub-intervals in the concerned sector. Conversely, if net improvements and net deteriorations in the sectoral trade balance appear in different sub-intervals, then no generalized relationship can be derived between the MIIT index for the aggregate time interval and its sub-intervals (Brulhart, 2000). As there are no established theoretical or empirical norms to verify whether a particular MIIT measure would be best, the empirical results on industrial adjustment may be sensitive to the choice of index as well as time interval and aggregation levels.

Literature Review

Global Evidence

The dynamic MIIT index proposed by Brulhart (1994) motivated a rich branch of literature on IIT adjustment costs. According to Brulhart, the plausibility of labour mobility being higher within than between industries is more evident if skills requirements are similar within IIT-intensive industries. Focusing on trade and economic adjustment in Irish chemical sector for three-digit SITC groups, Brulhart (1994) found that MIIT is, first, related to inter-industry trade adjustment and, second, to structural variables. Subsequently, a number of studies attempted to compute MIIT index both for developed (Cabral & Silva, 2006; Cernosa, 2012; Erlat & Erlat, 2003; Ferto, 2009; Lee & Sohn, 2004) and developing countries (Brulhart & Thorpe, 2000; Oliveras & Terra, 1997; Veeramani, 2007).

A branch of literature focused on a quality-based methodological approach for examining the simultaneous exports and imports of quality-differentiated goods by decomposing the IIT in HIIT with VIIT, based on Abd-el-Rahman (1991) framework. Extending the existing B index, Thom and McDowell (1999) proposed a method to distinguish between HIIT and VIIT at sectoral level. The analysis noted the prominence of IIT over inter-industry trade, with HIIT dominating VIIT, in case of Czech and Slovak Republics and Hungary. Considering two-digit sectoral data for Turkey, Erlat and Erlat (2003) decomposed the MIIT measured through B index into VIIT and HIIT components. They reported high MIIT, with HIIT dominating VIIT, and adjustment costs being predominantly of intra-industry type in nature. Drawing evidence from Portuguese and UK context, Cabral and Silva (2006) also noted that SAH may be more applicable to sectors characterized by horizontal than vertical IIT.

Over the years, empirical verification of the SAH to analyse the relationship between adjustments costs and different control variables emerged as a major area of research (Brulhart & Elliot, 1998; Brulhart & Thorpe, 2000; Ferto, 2009; Sarris et al., 1999; Tharakan & Calfat, 1999; Thorpe & Leitao, 2012). Several studies have used industry-level employment and industry’s total job turnover as a proxy for labour market adjustment process. Empirical studies based on job turnover as a dependent variable are based on the assumption that this captures adjustment costs. The argument is supported by the fact that changing jobs within the same sector implies fewer adjustment requirements than changing between sectors (Brulhart & Elliot, 1998). Cabral and Silva (2006) estimated a fixed-effects model linking the adjustment cost variable with changes in MIIT for Portugal and UK. The results pointed to lower adjustment costs being associated with IIT-type vis-a-vis inter-industry trade expansion. Based on individual-level data on manufacturing sector in UK, Brulhart et al. (2006) concluded that presence of IIT reduces workers movement both between the occupations and between industries in a stipulated time. This made a significant contribution to the literature as previous SAH studies ignored worker movements across occupations. Focusing on Hungary, Ferto (2009) confirmed that industry-specific variables may have a significant effect on employment changes and the associated adjustment costs. Analysing Australian manufacturing industries, Thorpe and Lietao (2012) reported that the relationship between changes in employment and increased IIT indices supports the SAH proposition.

Several studies, involving both developed and developing countries, however, failed to notice SAH predictions. Haynes et al. (2002) observed that labour transfers between sectors in UK, maintaining the same occupation/job, imply lesser adjustment costs than labour transfers within the same sector involving job requirement change. The analysis of Cernosa (2012) involving bilateral trade between Central European countries also did not support prevalence of SAH. The empirical analysis of Erlat and Erlat (2006), involving increase in IIT and adjustment costs due to trade expansion, did not observe SAH phenomenon in Turkey. Econometric analysis on Malaysia’s trade expansion during the high-growth period (1970–1994) reveals that IIT is related with relatively large payroll changes, which goes against the SAH predictions (Brulhart et al., 1999).

Indian Evidence

Several studies have attempted to analyse India’s IIT pattern. In the pre-1991 days, India witnessed a modest IIT level. Considering the 1960–1980 period, Pant and Barua (1986) noted that in spite of rise in trade flows, India’s IIT indices changed only modestly in a few commodity groups. As both exports and imports increased through tariff and other policy reforms in the post-reform period, an increasing trend in IIT, both at aggregate and sectoral levels, has been noted (Aggarwal & Chakraborty, 2017, 2019; Burange & Chaddha, 2008; Burange et al., 2017; Chakraborty & Chakraborty, 2005).

A few studies have computed MIIT in the Indian context. Veeramani (2002) reported aggregate and bilateral IITs across 43 industry groups (HS two-digit) in Indian manufacturing sectors and 51 partner countries for 1995. The findings revealed that transport equipment, instruments and apparatus, and stone and cement recorded the lowest level of MIITs. Focusing on 42 industry groups (HS two-digit), Veeramani (2004) noted that MIIT has significantly increased during post-liberalization period and rising shares of the changes in trade flows are of IIT-type in nature. Analysing India’s trade with partners over 1987–1988 to 2005–2006, Kelkar and Burange (2016) reported a rise in MIIT in manufacturing segment and lower adjustment costs within industries.

There exists a rich literature on various dimensions of employment trends in Indian manufacturing sectors, including trade-employment linkage. The positive influence of trade liberalization on firm-level productivity has been a well-researched area (Goldberg et al., 2010; Topalova & Khandelwal, 2011). However, these studies do not explicitly focus on the labour market adjustments and their consequent impact on employment changes. It has been noted that the increasing foreign value added (FVA) intensity of domestic production is leading to rise in productivity with low employment growth (Mohanty & Saha, 2017). A few studies explored the interrelationship between trade and employment creation. Sen (2009) assessed the direct and indirect effects of international trade on manufacturing employment for 25 industries at the ISIC three-digit level during 1975–1999. Raj and Sasidharan (2015) analysed the impact of international trade on manufacturing employment over the period 1980–2005. Evidence of strong trade-induced employment growth has however not been witnessed by these studies. Das et al. (2014) studied the impact of international trade on labour demand and employment elasticity in aggregate as well as disaggregated industries using Annual Survey of Industries (ASI) database at NIC four-digit level for 1991–2010. However, the elasticity level varied significantly across sectors. The poor linkage between trade and employment in the Indian manufacturing sector has been explained by the skill-intensive nature of India’s exports, resulting in labour productivity growth and consequent shedding of excess labour at firm-level (Raj & Sen, 2012). On the contrary, growing manufacturing trade deficit in the recent period may lead to associated adjustment requirements, leading to job losses within industrial sectors (Chaudhuri, 2015).

It is observed that empirical analysis on sector-level labour market dynamics, particularly in relation to MIIT and SAH in the Indian context, is a relatively less researched area. The present analysis, therefore, explores the relationship between MIIT and employment change for seven select manufacturing product groups over 2001–2015.

Data, Indices and Model

Data

The present analysis pertains to manufacturing commodities based on ASI data, as released by Central Statistical Office (CSO), under Ministry of Statistics and Programme Implementation, Government of India (GoI, n.d.). The product groups under the four-digit National Industrial Classification (NIC) levels are considered as industry. The NIC classifications have undergone changes over the study period. First, NIC-1998 classification, based on International Standard Industrial Classification (ISIC) Rev. 3, was followed to classify economic activities of the factories in ASI data over 1998–1999 to 2003–2004. Second, NIC-2004 was developed on the basis of ISIC Rev. 3.1 and ASI data over 2004–2005 to 2007–2008 had been reported in this format. Finally, ASI data in NIC-2008 classification, developed on the basis of ISIC Rev. 4, is available from 2008–2009 onwards. All the employment-specific and some industry-related and variables for the period 2000–2015 are collected directly from the publications of ASI. Some industry-specific data has also been drawn from PROWESS database maintained by the Centre for Monitoring Indian Economy, which provides firm-level audited financial results (CMIE, n.d.).

The analysis first identifies the major sectors in the Indian context, which experience simultaneous export and imports. The goal is to select product groups which represent a significant share in India’s trade basket and also experience trade overlap. Seven major manufacturing product groups, namely: chemicals, leather and footwear, textiles and garments, iron and steel, base metals, electrical machinery and equipment and vehicles and auto-components are selected, which collectively account for more than 40 per cent of India’s export and import flows (Aggarwal & Chakraborty, 2017). The trade data, available in Harmonized System (HS) classification at HS four-digit level (i.e., tariff headings), are obtained from Trade Map database, maintained by International Trade Centre (ITC) (ITC, n.d.). For conducting the empirical analysis over 2001–2015, a concordance between the 56 industry codes (identified at NIC four-digit level) and the corresponding trade codes (reported in HS) has been developed by matching the product descriptions in HS and NIC classifications, which is reported in Annexure 1.

Cross-Sectional IIT and MIIT

The Brulhart (1994) index, a measure of dynamism in trade overlap, is used for the purpose of empirical analysis in the current context. The values of the IIT (computed by GL index) and MIIT (computed by B index) for the selected product groups at five different points during 2001–2018 are reported in Table 1. The IIT levels have shown an increasing trend in sectors characterized by relatively lower capital-intensity, while a fluctuating trend has been noticed for machinery and vehicles sectors. It may further be seen that the levels of marginal IIT both at the aggregate and sectoral levels have shown fluctuations over 2001–2018, though an upward rising trend can be noted. The computed indices indicate that barring annual fluctuations, in all sectors, the export and import changes are coming closer over the years. In other words, with shares of the changes in trade flows to be of intra-industry type, the phenomenon of IIT in India is dynamic in nature.

The scatter plots between the MIIT and IIT variable, measured first between the B and GL index and then by B and the first difference of GL, that is, d(GL), for the sample period (56 industries, 2001–2015) are reported in Annexure 2. It is clearly noted from the reported R² values that B index is not strongly related to the IIT index both at the level and first difference. In other words, the distinction between MIIT and IIT in the Indian context seems relevant both conceptually and empirically.

Note on Employment Scenario

Share of employment in selected sectors (expressed as a percentage of overall employment in industrial sector) and their corresponding compound annual growth rates (CAGR) during 2001–2015 are shown in Table 2. It is observed that the selected industries in the study period account for more than 50 per cent of total employment across the manufacturing sectors. While the employment shares of chemical and textile and clothing sectors have declined, there has been a rise in the corresponding figures in leather and footwear, iron and steel and vehicle and auto-components. The employment shares in capital-intensive sectors of base metals and machinery and equipment have remained almost the same. The average CAGR of total employment has been reported for three periods, namely: 2001–2005, 2006–2010 and 2011–2015, which stand at 4.13, 5.29 and 1.58 per cent, respectively. A dampening trend in employment formation in Indian industrial sector is clearly observed in recent period, in line with growing casualization of the workforce as well as sectoral dynamics (Goldar & Aggarwal, 2010; Mehrotra & Parida, 2019). At the sectoral level, a decline in employment CAGR can be observed in chemicals, iron and steel and electrical machinery and equipment sectors during 2011–2015, underlining the much-discussed jobless growth phenomenon in the country (Tejani, 2015; Thomas, 2013). There is a need to understand whether trade in general and the IIT patterns in particular, influence the sectoral employment dynamics.

Empirical Method and Empirical Model

Given the time period (2001–2015) and various factors influencing the labour market dynamics across Indian industries, a balanced panel data model has been estimated here. There is a need for controlling non-stationarity in the data; in order to avoid spurious results (Baltagi, 2005; Bagchi & Bhattacharyya, 2019; Pesaran, 2015). The Harris-Tzavalis Test (1999), which has a null of unit root versus an alternative with a single stationary value, is performed to detect the presence of unit root among the explanatory variables. All the variables used in the regression analysis are found to be stationary except unskilled wage rate. Therefore, the regression model uses first difference of the unskilled wage rate variable, so that the modified series becomes stationary. Also, standardized FDI variable (logarithmic transformation) has been used in the model. Table 3 reports the statistic for all the variables included in the model using Harris-Tzavalis Test.

In addition, the endogeneity check for the explanatory variables has been performed in the analysis using two-stage least squares method. It is observed that the Wald chi-square test statistic of 58.71 (Prob: 0.00) is statistically significant. The null hypothesis of the Durbin and Wu-Hausman tests is that the variable under consideration can be treated as exogenous. Durbin score of 0.4832 (Prob 0.487) and Wu-Hausman statistic is 0.4812 (Prob 0.488) are not significant, so the null hypothesis of exogeneity is not rejected. Therefore, it can be noted that explanatory variables used in the panel data analysis, such as unskilled wage rate, FDI, trade openness, ratio of skilled to unskilled workers, are not endogenous.

Table 2.

Employment Share and CAGR (%) Across Select Product Groups

Commodity Sections	HS Codes	Employment Share				Employment CAGR (%)
Commodity Sections	HS Codes	2000	2005	2010	2015	2001–2005	2006–2010	2011–2015
Chemicals	28, 29	2.75	1.87	1.78	1.81	−3.71	6.31	−0.03
Leather and footwear	41, 42, 64	1.73	1.91	2.31	2.66	4.43	13.86	5.65
Textile and garments	50–63	20.59	20.90	18.94	19.09	5.67	−0.03	2.58
Iron and steel	72, 73	8.55	8.66	12.04	10.21	5.21	12.31	−2.30
Base metals	74–83	2.20	2.49	2.72	2.67	8.23	7.76	1.09
Electrical and electronics machinery and equipment	84, 85	9.88	9.64	10.81	9.69	3.88	9.45	−1.59
Vehicles and auto-components	87	4.82	5.60	7.08	8.27	8.27	10.80	4.19
Total		50.52	51.08	55.67	54.39	4.13	5.29	1.58

Source: The authors (computation from Annual Survey of Industries, GoI [n.d.]).

Table 3.

Harris-Tzavalis-Type Panel Unit Root Test Statistic

Variables	Rho	Z
${\| Δ E M P L \|}_{i t}$	0.3330	−18.6373***
${\| Δ E M P L \|}_{i t - 1}$	0.3259	−18.9154***
MIIT	0.0070	−31.3130***
MIIT*RGR	0.2844	−20.5266***
MIIT*(S/U)	−0.0097	−31.9599***
TO	0.0421	−29.9483***
(S/U)	−0.0570	−33.8011***
ΔWAGE	−0.2293	−40.4976***
CR	0.5480	−10.2834***
LFDI	0.5871	−8.7614 ***
PROD	0.5540	−10.0498***
PROD*LFDI	0.5901	−8.6448***
$(P R O D * (\frac{S}{U}))$	0.4983	−12.4298 ***
MIIT*LFDI	0.4816	−13.0176***
MIIT*PROD	0.4218	−13.6280***

Source: The authors (estimation using Stata: Release 14).

Notes: *** denotes the statistical significance at 0.01.

In line with the existing literature, the following panel data model is estimated to explore the relationship between the absolute value of total employment changes in Indian industries and the factors influencing the labour market dynamics.

\begin{matrix} {| Δ E M P L |}_{i t} = α_{0} + β_{1} {| Δ E M P L |}_{i t - 1} + β_{2} M I I T_{i t} + β_{3} {(M I I T * R G R)}_{i t} \\ \begin{array}{l} \begin{array}{l} + β_{4} (_{M I I T * (\frac{S}{U})) i t} + β_{5} T O_{i t} + β_{6} {(\frac{S}{U})}_{i t}) + β_{7} Δ W A G E_{i t} + \end{array} \\ \begin{matrix} β_{8} C R_{i t} + β_{9} L F D I_{i t} + β_{10} P R O D_{i t} + β_{11} {(P R O D * L F D I)}_{i t} + \\ β_{12} {(P R O D * (\frac{S}{U}))}_{i t} + β_{13} {(M I I T * L F D I)}_{i t} + \\ β_{14} {(M I I T * P R O D)}_{i t} + Y e a r ​ D_{t} + S e c t o r D_{i} + ε_{i t} \end{matrix} \end{array} \end{matrix}

(2)

where, α represents the constant term βs are coefficients L represents logarithmic transformation of the variables Δ represents absolute change of the variables EMPL_it represents employment for sector i in year t EMPL_it-1 represents employment for sector i in year (t−1) MIIT_it represents the marginal intra industry trade (MIIT) for sector i in year t (MIIT *RGR)_it represents the interaction term between MIIT and relative growth rate for sector i in year t (MIIT *(S/U))_it represents the interaction term between MIIT and ratio of skilled workers to unskilled workers for sector i in year t TO_it represents imports plus exports as a share of output for sector i in year t (S/U)_it represents ratio of skilled workers to unskilled workers for sector i in year t ΔWAGE_it represents change in unskilled wage rate for sector i in year t CR_it represents concentration ratio for sector i in year t FDI_it represents foreign direct investment inflows in sector i in year t PROD_it represents average labour productivity for sector i in year t (PROD*LFDI)_it represents the interaction term between labour productivity and logarithmic tranfortation of foreign direct investment for sector i in year t ${(P R O D * (\frac{S}{U}))}_{i t}$ represents the interaction term between average labour productivity and ratio of skilled workers to unskilled workers for sector i in year t ${(M I I T * L F D I)}_{i t}$ represents the interaction term between MIIT and logarithmic tranfortation of foreign direct investment for sector i in year t ${(M I I T * P R O D)}_{i t}$ represents the interaction term between MIIT and average labour productivity for sector i in year t Year D_t represents year dummy variable Sector D_i represents sector dummy variable

A dynamic AR model is estimated here since the independent variables in time t include the value of the dependent variable during the previous time period, that is, the lagged response ${| Δ E M P L |}_{i t - 1}$ in a way that is not explained by the other factors. The current analysis estimates a Generalized Method of Moments (GMM) model to alleviate the bias resulting from using simple panel data methodology. All the variables used in the analysis are stationary. The description of the variables used in the empirical analysis and the corresponding data sources are summarized in Annexure 3.

The dependent variable ${| Δ E m p l |}_{i t}$ , absolute change in the employment in the ith industry in the tth period, is a proxy for adjustment costs (Brulhart, 1999). The present analysis takes note of the existing literature on whether the lagged differences (i.e., ${| Δ E m p l |}_{i (t - 1)}$ ) are valid instruments for the level in an autoregressive model (Blundell & Bond, 1998; Bond, 2002). The rationale for this procedure is twofold (Ferto, 2009). First, given the dynamic nature of employment changes, past changes of the series may influence current variations. Second, the use of a lagged dependent variable can provide deeper insight into short- and long-run adjustment effects. Following Thorpe and Leitao (2012), the variable is defined as follows:

Δ E M P L = 2 (E M P L_{t} - E M P L_{t - 1}) / (E M P L_{t} + E M P L_{t - 1})

(3)

A set of trade-related control variables have been included in the model in line with existing literature. According to SAH literature, it is expected that |ΔEMPL| will have a negative relationship to MIIT (Brulhart, 1999; Brulhart & Elliott, 1999). In other words, the sectors characterized by high MIIT would experience lesser adjustment costs in terms of absolute changes in employment levels (Thorpe & Leitao, 2012). As from ASI sources, the data on inter-sectoral movement of labour in India is not obtained, absolute change in employment level across sectors is considered as a dependent variable in the current context. In the current analysis, Trade Openness (TO) of the sector have been included as a key independent variable, as greater openness is expected to significantly influence the labour market adjustments for the given industry. Higher exposure to trade implies stronger competitive pressure, resulting in higher necessity of firms and industries to adapt more frequently to evolving competitive positions (Brulhart, 2000; Lee & Sohn, 2004).

A set of labour-related control variables have also been included. The current analysis incorporated ratio of skilled to unskilled workers (S/U) in the model, as rising skill-intensity among workers may reflect higher absolute changes in employment. Worker’s resistance to move to a different industry is likely to be higher if they originate in a high-wage industry, especially among the high-skilled workers. Conversely, an unskilled worker may be more amenable to the idea due to low adjustment costs in the specified sector in comparison to his high-skilled counterpart (Cabral & Silva, 2006). Moreover, wage rate is expected to have a negative influence on employment changes (Brulhart & Elliott, 2002; Brulhart et al., 2006). In the current context, given the possible substitution of unskilled workers with technical progress and industrial adjustment (Kapoor, 2016), unskilled wage rate prevailing in an industry is considered as an independent variable.

The model also includes a set of industry-related control variables. A positive relationship between employment changes and market concentration is expected, as lop-sided market dominance will entail relatively higher intra-sectoral employment reallocations (Rijesh, 2019). Labour productivity is expected to be positively associated with employment changes (Brulhart & Thorpe, 2000; Ferto, 2009; Ferto & Soos, 2010), especially in an expanding industry characterized by economies of scale (Rijesh, 2019). The relationship between sectoral FDI inflows and corresponding adjustment costs may, however, be dubious. If FDI is concentrated in labour-intensive industries, a positive impact on level of employment is expected (Ghosh & Sinha Roy, 2015). Alternatively, investments by the multinational enterprises in highly capital-intensive industries are expected to have lower employment elasticity of output as compared to domestic firms adopting labour-intensive techniques of production (Pradhan et al., 2004).

Finally, the present analysis proposes six important interaction terms in the regression model. For the first four terms, the idea is to note how marginal intra industry trade (MIIT) may behave in association with other key industry-specific variables. First, an interaction term between MIIT index and relative growth rate (i.e., ratio of ith industry’s output growth rate to aggregate output growth rate of the industrial sector) has been incorporated. A positive relationship for this interaction variable (MIIT*RGR) is expected, that is, a sector witnessing relatively higher growth and MIIT, may employ more workers and experience greater labour market adjustment and vice versa. Second, an interaction term between MIIT index and skilled worker intensity in industry, namely, MIIT*(S/U), has been included. A positive coefficient is expected as sectors characterized by rising skill intensity and MIIT may employ more workers and experience greater labour market adjustment and vice versa. Third, an interaction between MIIT index and incremental FDI inflows, namely (MIIT*LFDI) has been incorporated. The sign of the coefficient may, however, be dubious, depending on the relationship between the objective behind incremental FDI inflows (i.e., export-driven or domestic market capturing) and corresponding trade pattern. Fourth, an interaction between MIIT index and average labour productivity, that is, (MIIT*PROD) has been included in the analysis. A positive coefficient is expected as sectors characterized by rising labour productivity and MIIT may employ more workers necessitating greater labour market adjustment and vice versa. Fifth, an interaction term between average labour productivity and incremental FDI inflows has been incorporated. In Indian context, FDI positively influences productivity (Pradeep et al., 2017). The sign of the coefficient of the interaction term may, however, be dubious, depending on the relationship between the nature of incremental FDI inflows and corresponding adjustment costs. For instance, FDI in highly capital-intensive industries is expected to lead to lower employment elasticity of output even in the productive sectors of the economy. Finally, an interaction term between average labour productivity and skilled worker intensity in an industry, namely (PROD*(S/U)) has been included. A positive coefficient for this interaction term is expected, as sectors characterized by relatively higher average labour productivity associated with higher skill-intensity, would entail higher adjustment costs in the labour market.

Empirical Results

The summary statistics for the variables selected for the empirical analysis is provided in Table 4. The current analysis then estimates a ‘Generalized Method of Moments’ (GMM) model to alleviate the bias resulting from using simple panel data methodology.

The empirical estimates in a dynamic framework for India’s industry-level adjustment effects are summarized in Table 5. The Hansen’s J test statistic is reported as it provides a test of over-identifying restrictions, that is, a test of the null hypothesis that the instrument set is appropriate for the data. Because the p-value is greater than 0.1 as shown in Table 5, therefore, the null hypothesis is not rejected, indicating that appropriate set of instrumental variables are used. The Sargan test statistic has also been reported. Thus, the analysis takes into account that instrumental variables are uncorrelated with errors.

Several conclusions emerge from the empirical results. The coefficients of the lagged employment changes ${| Δ E M P L |}_{i (t - 1)}$ is positive and significant for all the model specifications, indicating that the past values of changes in employment level influence the labour market adjustment process in the current period.

The trade-specific determinants are noted next. First, the relationship between MIIT (B Index) and absolute change in employment is positive and significant. The result implies that sectors characterized by higher MIIT levels are witnessing greater employment adjustments. The obtained results do not conform to the SAH predictions in the Indian context. Second, the coefficient of TO is positive and significant for all model specifications, indicating that greater trade openness leads to higher labour market adjustments. On the whole, in line with theoretical predictions, trade dynamics leaves a significant influence on labour market adjustments.

The labour-specific determinants are discussed next. First, the coefficients of (S/U) are positive and significant for all specifications, that is, a larger relative presence of skilled workers vis-à-vis unskilled ones is associated with greater absolute employment changes. It has been observed that Indian exports from skill-intensive sectors (e.g., leather and footwear) are on the rise, which may lead to scale expansion and consequent rise in labour demand. Second, the coefficient of wage for unskilled workers (WAGE) is found to be negative and significant, indicating that higher increase in unskilled wages may escalate the cost to the firms, thereby generating a lower demand for labour, in turn leading to decline in employment levels and vice versa. While there can be a direct rise in demand for unskilled workers if the wages fall, an indirect demand for skilled workers may also simultaneously arise, reflected in adjustments in overall employment in the opposite direction. Third, labour productivity (PROD) on the other hand is positively associated with employment changes, signifying greater employment rise in more productive sectors, putting them on a growth path and vice versa. It can be noted that the labour market-related factors play a key role in the adjustment process.

Table 4.

Summary Statistics

Variable	Observation	Mean	Std. Dev.	Min	Max
ΔEMPL(t)	840	0.150	0.203	0.000	1.848
ΔEMPL(t−1)	840	0.144	0.202	0.000	1.848
MIIT	840	0.332	0.325	0.000	0.999
MIIT*RGR	840	0.012	6.970	−111.961	87.008
MIIT* (S/U)	840	0.120	0.181	0.000	3.849
TB	840	0.400	0.490	0	1
TO	840	1.219	4.669	0.001	98.890
S/U	840	0.369	0.416	0.121	10.738
ΔWAGE	840	0.073	0.190	−1.874	1.968
CR	840	0.250	0.113	0.145	0.719
LFDI	840	10.173	0.708	6.544	11.424
PROD	840	4.896	4.127	−0.568	26.336
PROD * LFDI	840	51.081	44.435	−5.187	287.771
$(P R O D * (\frac{S}{U}))$	840	1.897	1.895	−0.144	13.640
MIIT*LFDI	840	3.388	3.346	0	10.817
MIIT*PROD	840	1.623	2.438	−0.488	21.696

Source: The authors

Table 5.

Regression Results on Determinants of Absolute Change in Employment

Independent Variables	Dependent Variable: \|ΔEmpl(t)\|
Independent Variables	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 9
\|ΔEmpl(t−1)\|	0.428***	0.428***	0.426***	0.433***	0.409***	0.411***	0.425***	0.432***	0.419***
	(0.005)	(0.011)	(0.003)	(0.006)	(0.011)	(0.016)	(0.010)	(0.009)	(0.009)
TO	0.004***	0.005***	0.006***	0.005***	0.004***	0.005***	0.005***	0.005***	0.005***
	(0.001)	(0.001)	(0.001)	(0.001)	(0.001)	(0.002)	(0.001)	(0.001)	(0.001)
S/U	0.001	0.013**	0.024*
	( 0.007)	(0.006)	(0.014)
MIIT		0.078***	0.080***
		(0.005)	(0.006)
ΔWAGE			−0.065***	−0.067***	−0.042***	−0.089***	−0.080***	−0.099***	−0.072***
			(0.010)	(0.012)	(0.010)	(0.015)	(0.009)	(0.019)	(0.020)
MIIT*(S/U)				0.098***
				(0.019)
MIIT*RGR					0.001***	0.001***	0.001***

					(0.001)	(0.001)	(0.001)
CR					0.285***	0.422***	0.421***	0.418***	0.282***
					(0.039)	(0.035)	(0.035)	(0.055)	(0.040)
LFDI					0.005
					(0.005)
PROD						0.007***
						(0.001)
PROD*LFDI							0.001***
							(0.001)
PROD*(S/U)								0.019***
								(0.002)
MIIT*LFDI								0.008***
								(0.001)
MIIT*PROD									0.007***
									(0.001)
Year dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Sector dummies	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
N	840	840	840	840	840	840	840	840	840
Sargan test statistic	55.96	55.41	53.41	53.77	53.02	52.45	51.97	52.32	52.08
Hansen’s J chi-square statistic	0.802	0.861	0.826	0.812	0.821	0.854	0.864	0.819	0.856
P-value	0.454	0.347	0.385	0.435	0.394	0.355	0.346	0.407	0.350

Source: The authors

Notes: Figure in the parenthesis shows the standard errors of the estimated coefficient. ***, ** and * imply estimated coefficient is significant at 0.01, 0.05 and 0.10 level, respectively.

Next, the sector-specific determinants are discussed. First, the coefficient of industrial concentration (CR) is found to be positively associated with employment changes. An exploratory analysis of raw data reveals that the four-firm concentration is generally declining, and transition of market structure is more likely to entail relatively higher intra-sectoral employment reallocation. Second, the coefficient of incremental FDI inflow (LFDI) is positive but non-significant. The result may indicate that foreign investment and technology inflow might target the growing domestic segments (e.g., textile and garments) without much technology transfer, leading to higher employment elasticity of output in labour-intensive segments. However, a strong relationship in a systemic fashion is not there.

Lastly, all the interaction terms incorporated in the model, namely: (MIIT*RGR), (MIIT*(S/U)), (MIIT*LFDI), (MIIT*PROD), (PROD*LFDI) and (PROD*(S/U)), are found to be positive and significant. The implications of these results in the Indian context are the following. First, high relative growth rate in an industry which also witness rise in IIT-type trade, may lead the firms therein to expand and employ more productive and competitive resources, including labour. Second, sectors characterized by higher MIIT and skill-intensity are witnessing growth, leading to higher employment adjustments. Thirdly, industries characterized by higher MIIT and incremental FDI inflows lead to greater labour market adjustment. Fourthly, more productive industries characterized by higher MIIT experience greater labour market adjustments. The results indicate that the positive relationship between MIIT and absolute employment adjustments noted earlier can be explained by favourable industry-specific undercurrents. Fifthly, industries characterized by high labour productivity and higher incremental FDI inflows lead to greater labour market adjustments, given the growth impetus within an expanding industry and vice versa. Finally, rise in productivity in the skill-intensive sectors may enable firms to employ more skilled workers and enjoy economies of scale, leading to absolute changes in employment. It is therefore observed that the influences of the industry-level variables and their interactions on labour market adjustment are along anticipated lines.

Conclusion

Over the last two decades, India has witnessed a series of reforms, namely: reduced import tariff, easing of business norms, improved existing infrastructure, etc., so as to develop as a major hub for production and exports through rising industrial efficiency. On the other hand, the entry of foreign firms in the country has been facilitated to ensure greater FDI inflows and technology transfer, with a goal to augment both volume and quality of exports. This was expected to spread the technology spillover effect across Indian firms, thereby hastening the process of integrating them with the international production networks (IPNs). The participation of the country in multiple regional trade agreements (RTAs) since 2005, launch of the recent initiatives like Make-in-India (2014), Act East Policy (2014), Skill India Mission (2015), etc., are conscious moves to expedite the process further.

The increasing internationalization of Indian industrial firms coupled with the growing presence of foreign players in the country leads to the growth of simultaneous export and import within same product categories (involving trade in intermediates, parts and components and final products). The evolving changes enabled the firms to specialize in and consequently export narrower product lines on one hand, while importing quality inputs and becoming part of the upstream value chains of foreign players on the other. The resulting rise in IIT, however, influences various segments of the industrial sector differently. For India, a densely populated developing country, the influence of the resulting IIT on labour market adjustment, particularly job losses, can emerge as a serious policy concern. In this larger context, the present analysis explored the relevance of the SAH phenomenon in select industrial sectors in India, that is, whether trade expansion in presence of dynamic IIT (measured by MIIT), entails better labour market stability.

The empirical results indicate that several trade-related, labour-oriented and industry-specific variables have a significant effect on absolute changes in employment levels. MIIT, increase in productivity, skilled workforce intensity, industrial concentration, incremental FDI inflows and trade openness positively influence absolute employment changes, whereas unskilled wage exerts a negative impact on the same. The obtained results do not strictly support the SAH predictions in the Indian context. The analysis further concludes that high relative growth rate, skill intensity, incremental FDI inflows and average labour productivity in a sector, also characterized by higher MIIT, may lead the firms to employ more productive and competitive resources, resulting in higher absolute changes in employment. Hence deepening India’s participation in IPNs and facilitation of FDI inflows may not automatically translate into a beneficial outcome. In fact, India’s recent pull-out from the RCEP negotiations, citing economic interests and national priorities (GoI, 2019), can be closely explained by the widening trade deficit in several sectors, characterized by lower skill-intensity and poorer productivity. The launch of the Atmanirbhar Bharat Abhiyan (2020) in May 2020 can be treated as a conscious attempt to lower import-dependence in manufacturing segments. The policymakers, therefore, need to facilitate skill-augmentation and productivity rise as well as technology transfer through the appropriate policy measures under ‘Make-in-India’, ‘Skill-India’ and other initiatives. Only then the necessary export impetus, upstream ‘domestic’ value chain integration and the associated capacity to absorb more workers in various tiers of the industrial sector can be generated.

A limitation of the present analysis is that due to the paucity of data on inter-sectoral movements of labour, it has not been possible to evaluate overall reallocation adjustment effects of IIT in the aggregate industrial sector. Future research, therefore, could focus on that aspect, if time-series data on inter-sectoral movements of labour is obtained across industrial classifications for India.

Footnotes

Acknowledgements

The helpful comments and suggestions from two anonymous reviewers on an earlier draft of the article are sincerely appreciated. The authors are, however, responsible for any remaining errors.

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.

Appendix

Annexure 3.

Variable Description and Source of Data Used in the Empirical Model

Sl. No.	Variable	Variable Description	Data Source
1	ΔEMPL	Computed with No. of Persons Employed data obtained from Annual Survey of Industries (ASI), GoI (n.d.) by following Thorpe and Leitao (2012), as expressed in equation (3).	Computed by author
2	TO	Computed by expressing total trade (export plus import) as a share of total output in an industry. The export and import data are obtained from Trade Map, ITC (n.d.). The total output data has been taken from ASI, GoI (n.d.).	Computed by author
3	MIIT	Computed with import and export data across sectors obtained from Trade Map, ITC (n.d.), by following Brulhart (1994), as expressed in equation (1).	Computed by author
4	MIIT*RGR	Multiplication of MIIT Index and Relative Growth Rate. Relative Growth Rate is computed by dividing growth rate of Total Output in each sector by growth rate of Total Output in all the sectors. Data obtained from ASI, GoI (n.d.).	Computed by author
5	S/U	Computed by dividing number of skilled and unskilled worker data taken from ASI, GoI (n.d.) at each NIC four-digit level.	Computed by author
6	MIIT* (S/U)	Multiplication of MIIT Index and S/U ratio. S/U ratio computed from ASI, GoI (n.d.) data.	Computed by author
7	CR	Computed by calculating Herfindahl-Hirschman Index (HHI) at the sectoral level. HHI is calculated by squaring the market share of each firm competing in a market, summing the resulting numbers and then taking the square root of the total summation. For computing market share, firm-level Sales data has been taken from Prowess database (CMIE, n.d.).	Computed by author
8	ΔWAGE	Unskilled wage rate computed by dividing the monetary value of unskilled wage bill by the number of unskilled workers at each NIC four-digit level, as obtained from ASI, GoI (n.d.).	Computed by author
9	LFDI	Sector-wise FDI inflows obtained from SIA Statistics, Department for Promotion of Industry and Internal Trade (DIPP, n.d.).	Compiled by author
10	PROD	Computed by dividing the Net Value Added by total employment in a given industry. Data obtained from ASI, GoI (n.d.).	Computed by author
11	PROD*LFDI	Multiplication of Average Labour Productivity and value of incremental Foreign Direct Investment Inflows. Data obtained from ASI, GoI (n.d.) and DIPP (n.d.), respectively.	Computed by author
12	PROD*(S/U)	Multiplication of Average Labour Productivity and ratio of skilled to unskilled workers. Data obtained for computation of both series from ASI, GoI (n.d.).	Computed by author
13	MIIT*LFDI	Multiplication of MIIT Index and incremental FDI Inflows. Data obtained for computation of both series from ITC (n.d.) and DIPP (n.d.), respectively.	Computed by author
14	MIIT*PROD	Multiplication of MIIT Index and Average Labour Productivity. Data obtained for computation of both series from ITC (n.d.) and DIPP (n.d.), respectively.	Computed by author

Source: Authors construction.

Note: First difference of the variable WAGE (ΔWAGE) has been incorporated in the analysis so that the modified series is stationary. Rest of the variables are stationary and hence incorporated in the model without transformation.

References

Abd-el-Rahman

K. S.

(1991). Firms competitive and national comparative advantages as joint determinant of trade composition. Weltwirtschaftliches Archiv, 127(1), 83–97.

Aggarwal

, & Chakraborty

(2017). Determinants of India’s bilateral intra-industry trade over 2001–2015: Empirical results. South Asia Economic Journal, 18(2), 296–313.

Aggarwal

, & Chakraborty

(2019). Which factors influence India’s intra-industry trade? Empirical findings for select sectors. Global Business Review. https://journals.sagepub.com/doi/10.1177/0972150919868343

Aturupane

, Djankov

, & Hoekman

(1999). Horizontal and vertical intra-industry trade between Eastern Europe and the European Union. Review of World Economics, 135(1), 62–81.

Azhar

A. K. M.

, & Elliott

R. J. R.

(2004). Economic integration and evolution of trade: A geometric interpretation of trade measures. Journal of Economic Integration, 19(4), 651–666.

Azhar

A. K. M.

, & Elliott

R. J. R.

(2008). On the measurement of changes in product quality in marginal intra-industry trade. Review of World Economics, 144(2), 225–247.

Bacchetta

, & Dellas

(1997). Firm restructuring and the optimal speed of trade reform. Oxford Economic Papers, 49(2), 291–306.

Bagchi

, & Bhattacharyya

(2019). Country-specific determinants of intra-industry trade in India. Foreign Trade Review, 54(3), 129–158.

Balassa

(1966). Tariff reductions and trade in manufacturers among the industrial countries. American Economic Review, 56(3), 466–473.

10.

Baldwin

R. E.

, Mutti

J. H.

, & Richardson

(1980). Welfare effects on the United States of a significant multilateral tariff reduction. Journal of International Economics, 10(3), 405–423.

11.

Baltagi

B. H.

(2005). Econometric analysis of panel data. John Wiley & Sons.

12.

Banerjee

, & Bhattacharyya

(2004). Does economic development cause intra-industry trade? The case of India: 1971 to 2000. Arthaniti: Journal of Economic Theory and Practice, 3(1–2), 30–48.

13.

Blundell

, & Bond

(1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics, 87(1), 115–143.

14.

Bond

, (2002). Dynamic panel data models: A guide to micro data methods and practice. Portuguese Economic Journal, 1(2), 141–162.

15.

Bown

C. P.

, & Irwin

D. A.

(2016). The GATT’s starting point: Tariff levels circa 1947 (Policy Research Working Paper 7649). World Bank.

16.

Brulhart

(1994). Marginal intra-industry trade: Measurement and relevance for pattern of industrial adjustment. Weltwirtschaftliches Archiv, 130(3), 600–613.

17.

Brulhart

(1999). Marginal intra-industry trade and trade-induced adjustment: A survey. In Brülhart

& Hine

R. C.

(Eds.), Intra-industry trade and adjustment: The European experience. Macmillan.

18.

Brulhart

(2000). Dynamics of intraindustry trade and labor-market adjustment. Review of International Economics, 8(3), 420–435.

19.

Brulhart

, & Elliott

R. J. R.

(1998). Adjustment to the European single market: Inferences from intra-industry trade patterns. Journal of Economic Studies, 25(3), 225–247.

20.

Brulhart

, & Elliott

R. J. R.

(2002). Labour-market effects of intra-industry trade: Evidence for the United Kingdom. Weltwirtschaftliches Archiv, 138(2), 207–228.

21.

Brulhart

, Elliott

R. J. R.

, & Lindley

(2006). Intra-industry trade and labour-market adjustment: A reassessment using data on individual workers. Review of World Economics, 142(3), 521–545.

22.

Brulhart

, & Thorpe

(2000). Intra-industry trade and adjustment in Malaysia: Puzzling evidence. Applied Economics Letters, 7(11), 729–733.

23.

Burange

L. G.

, & Chaddha

S. J.

(2008). Growth in India’s intra-industry trade (Working Paper UDE 24/2/2008). Department of Economics, University of Mumbai.

24.

Burange

L. G.

, Thakur

, & Kelkar

H. K.

(2017). Foreign direct investment and intra-industry trade in India’s manufacturing sector: A causal relationship. Foreign Trade Review, 52(4), 203–218.

25.

Cabral

, & Silva

(2006). Intra-industry trade expansion and employment reallocation between sectors and occupations. Review of World Economics, 142(3), 496–520.

26.

Centre for Monitoring Indian Economy (n.d.). Prowess database.

27.

Cernosa

(2012). Central European countries and the smooth adjustment hypothesis. Journal of Contemporary Issues in Economics and Business, 58(5–6), 14–22.

28.

Chakraborty

, & Chakraborty

(2005). Indian exports in the post-transitory phase of WTO: Some exploratory results and future concerns. Foreign Trade Review, 40(1), 3–26.

29.

Chaudhuri

(2015). Import liberalisation and premature deindustrialisation in India. Economic and Political Weekly, 50(4), 60–69.

30.

Das

, Raychaudhuri

, & Sinha Roy

(2014). Trade and labor demand in an emerging market economy: An analysis for Indian manufacturing during 1991–2010. The Journal of Developing Areas, 48(3), 353–379.

31.

Datt

(1994). Jobless growth: Implications of new economic policies. Indian Journal of Industrial Relations, 29(4), 407–427.

32.

Deb

A. K.

, & Ray

S. C.

(2014). Total factor productivity growth in Indian manufacturing: A biennial Malmquist analysis of inter-state data. Indian Economic Review, 49(1), 1–25.

33.

Department of Industrial Policy and Promotion (DIPP) (n.d.). Secretariat of Industrial Assistance (SIA) statistics. https://dipp.gov.in/publication/sia-statistics

34.

Devadason

E. S.

(2012). Enhancing China–India trade cooperation: Complementary interactions? The China Review, 12(2), 59–84.

35.

Dhar

(2019). India’s withdrawal from the regional comprehensive economic partnership. Economic & Political Weekly, 54(45), 59–65.

36.

Dickey

D. A.

, & Fuller

W. A.

(1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association, 74(366), 427–431.

37.

Drèze

(1960). Quelques Réflexions Sereines sur l’Adaptation de l’Industrie Belge au Marché Commun. Comptes-rendus des Travaux de la Société d’Economie Poli- tique de Belgique, 275, 3–37.

38.

Drèze

(1961). Les Exportations Intra-CEE en 1958 et la Position Belge. Recherches Economiques de Louvain, 27(8), 717–738.

39.

Erlat

, & Erlat

(2003). Measuring intra-industry trade and marginal intra-industry trade: The case for Turkey. Emerging Markets Finance and Trade, 39(6), 5–38.

40.

Erlat

, & Erlat

(2006). Intraindustry trade and labor market adjustment in Turkey. Another piece of puzzling evidence? Emerging Markets Finance and Trade, 42(5), 5–27.

41.

Ferto

(2009). Labour market adjustment and intra-industry trade: The effects of association on the Hungarian food industry. Journal of Agricultural Economics, 60(3), 668–681.

42.

Ferto

, & Soos

K. A.

(2010). Marginal intra-industry trade and adjustment cots in the first phase of transition: A Hungarian-Polish comparison. Journal of Economic Studies, 37(5), 495–504.

43.

Ghosh

, & Sinha Roy

(2015). FDI, technology acquisition and labour demand in an emerging market economy: A firm level exploration of Indian manufacturing industries. The Journal of Industrial Statistics, 4(1), 19–36.

44.

Globerman

, & Dean

J. W.

(1990). Recent trends in intra-industry trade and their implications for future trade liberalization, Weltwirtschaftliches Archiv, 126(1), 25–49.

45.

Goldar

, & Aggarwal

S. C.

(2010). Informalization of industrial labour in India: Are labour market rigidities and growing import competition to blame? https://www.isid.ac.in/~pu/conference/dec_10_conf/Papers/BishwanathGoldar.pdf

46.

Goldberg

P. K.

, Khandelwal

A. K.

, Pavcnik

, & Topalova

(2010). Multiproduct firms and product turnover in the developing world: Evidence from India. The Review of Economics and Statistics, 92(4), 1042–1049.

47.

Government of India (GoI) (2019). India exploring trade agreements with USA & EU; FTAs with Japan, Korea & ASEAN being reviewed; No trade agreements in a hurry says Piyush Goyal. Press Information Bureau. https://pib.gov.in/newsite/PrintRelease.aspx?relid=194281

48.

Government of India (GoI) 2020, ‘PM gives a clarion call for Atmanirbhar Bharat’, May 12, available at: https://pib.gov.in/PressReleseDetail.aspx?PRID=1623391 (accessed July 16, 2020).

49.

Government of India. (n.d.). Annual Survey of Industries Database. Ministry of Statistics and Programme Implementation, Central Statistical Organization. http://www.csoisw.gov.in/cms/en/1023-annual-survey-of-industries.aspx

50.

Greenaway

, & Milner

(1986). The economics of intra-industry trade. Blackwell.

51.

Greenaway

, Hine

R. C.

, & Milner

(1995). Vertical and horizontal intra-industry trade: A cross industry analysis for the United Kingdom. The Economic Journal, 105(433), 1505–1518.

52.

Greenaway

, Hine

R. C.

, Milner

, & Elliott

R. J. R.

(1994). Adjustment and the measurement of marginal intra-industry trade. Weltwirtschaftliches Archiv, 130(2), 418–427.

53.

Grubel

H. G.

(1967). Intra-industry specialization and the pattern of trade. Canadian Journal of Economics and Political Science, 33(3), 374–388.

54.

Grubel

H. G.

, & Lloyd

P. J.

(1971). The empirical measurement of intra-industry trade. Economic Record, 47(120), 494–517.

55.

Grubel

H. G.

, & Lloyd

P. J.

(1975). Intra-industry trade: The theory and measurement of international trade with differentiated products. Macmillan.

56.

Hamilton

, & Kniest

(1991). Trade liberalisation, structural adjustment and intra-industry trade. Weltwirtschaftliches Archiv, 127(2), 356–367.

57.

Haynes

, Upward

, & Wright

(2002). Estimating the wage costs of inter and intra-sectoral adjustment. Weltwirtschaftliches Archiv, 138(2), 229–253.

58.

International Trade Centre (n.d.). Trade Map database. http://www.trademap.org/Index.aspx

59.

Kapoor

(2016). Technology, jobs and inequality: Evidence from India’s manufacturing sector (Working Paper No. 313). Indian Council for Research on International Economic Relations.

60.

Kelkar

H. K.

, & Burange

L. G.

(2016). India’s vertical and horizontal intra-industry trade during post-liberalization period. In Chakraborty

& Mukherjee

(Eds.), Trade, investment and economic development in Asia: Empirical and policy issues. Routledge.

61.

Krugman

(1981). Intra-industry specialization and the gains from trade. Journal of Political Economy, 89(5), 959–973.

62.

Lee

, & Sohn

(2004). South Korea’s marginal intra-industry trade and the choice of preferential partners. Asian Economic Papers, 3(3), 94–116.

63.

Lovely

, & Nelson

(2000). Marginal intra-industry trade and labour adjustment. Review of International Economics, 8(3), 469–479.

64.

Lovely

, & Nelson

(2002). Intra-industry trade as an indicator of labor market adjustment. Weltwirtschaftliches Archiv, 138(2), 179–206.

65.

Mazumdar

, & Sarkar

(2004). Reforms and employment elasticity in organised manufacturing. Economic & Political Weekly, 39(27), 3017–3029.

66.

Mehrotra

, & Parida

J. K.

(2019). India’s employment crisis: Rising education levels and falling non-agricultural job growth (Working Paper 2019–04). Centre for Sustainable Employment, Azim Premji University.

67.

Melitz

M. J.

(2003). The impact of trade on intra-industry reallocations and aggregate industry productivity. Econometrica, 71(6), 1695–1725.

68.

Mohan

N. C.

(2014). Unemployment in an era of jobless growth (Working Paper No. 159). Institute for Studies in Industrial Development.

69.

Mohanty

S. K.

, & Saha

(2017, June 21). Employment generation in Indian manufacturing sector relevance of export sector in job creation [Paper presentation]. Workshop on Job Creation in Manufacturing Sector: Strategy for Sustainable Economic Growth in South Asia, New Delhi.

70.

Nag

(2016). Emerging production network between India and ASEAN: An analysis of value added trade in select industries. In Chakraborty

& Mukherjee

(Eds.), Trade, investment and economic development in Asia: Empirical and policy, issues (pp. 41–67). Routledge.

71.

Neary

J. P.

(1985). Theory and policy of adjustment in an open economy. In Greenaway

(Ed.), Current issues in international trade. Macmillan.

72.

Oliveras

, & Terra

(1997). Marginal intra-industry trade index: The period and aggregation choice. Weltwirtschaftliches Archiv, 133(1), 170–178.

73.

Panagariya

(2004). India in the 1980s and 1990s: A triumph of reforms (IMF Working Paper No. 43). International Monetary Fund.

74.

Pant

, & Barua

(1986). India’s intra-industry trade: 1960–1980 (Discussion Paper No. 8). International Trade and Development Division, School of International Studies, Jawaharlal Nehru University.

75.

Pesaran

M. H.

(2015). Time series and panel data econometrics. Oxford University Press.

76.

Phillips

P. C. B.

, & Perron

(1988). Testing for a unit root in time series regression. Biometrika, 75(2), 335–346.

77.

Pradhan

J. P.

, Abraham

, & Sahoo

M. K.

(2004). Foreign direct investment and labour: The case of Indian manufacturing. Labour and Development, 10(1), 58–79.

78.

Pradeep

, Bhattacharya

, & Chen

(2017). Spillover effects of research and development, exports and foreign investment on productivity: Empirical evidence from Indian manufacturing. Journal of South Asian Development, 12(1), 18–41.

79.

Raj

S. N. R.

, & Sasidharan

(2015). Impact of foreign trade on employment and wages in Indian manufacturing. South Asia Economic Journal, 16(2), 209–232.

80.

Raj

S. N. R.

, & Sen

(2012). The impact of international trade on manufacturing employment in India. In Pushpangadan

& Balasubramanyam

V. N.

(Eds.), Growth, development, and diversity: India’s record since liberalization. Oxford University Press.

81.

Rijesh

(2019). International trade and productivity growth in Indian industry: Evidence from the organized manufacturing sector. Journal of South Asian Development, 14(1), 1–39.

82.

Sarris

A. H.

, Papadimitrou

, & Mavrogiannis

(1999). ‘Greece’. In Brülhart

& Hine

R. C.

(Eds.), Intra-industry trade and adjustment: The European experience. Macmillan.

83.

Sen

(2009). International trade and manufacturing employment: Is India following the footsteps of Asia or Africa? Review of Development Economics, 13(4), 765–777.

84.

Tejani

(2015). Jobless growth in India: an investigation. Cambridge Journal of Economics, 40(3), 843–870.

85.

Tharakan

P. K.

, & German

M. C.

(1999). ‘Germany’. In Brülhart

& Hine

R. C.

(Eds.), Intra-industry trade and adjustment: The European experience. Macmillan.

86.

Thom

, & McDowell

(1999). Measuring marginal intra-industry trade. Weltwirtschaftliches Archiv, 135(1), 48–61.

87.

Thomas

J. J.

(2013). Explaining the ‘jobless’ growth in Indian manufacturing. Journal of the Asia Pacific Economy, 18(4), 673–692.

88.

Thorpe

, & Lietao

N. C.

(2012). Marginal intra-industry trade and adjustment costs: The Australian experience. Economic Papers, 31(1), 123–131.

89.

Topalova

, & Khandelwal

(2011). Trade liberalization and firm productivity: The case of India. Review of Economics and Statistics, 93(3), 995–1009.

90.

Veeramani

(2002). Intra-industry trade of India: Trends and country-specific factors. Weltwirtschaftliches Archiv, 138(3), 509–533.

91.

Veeramani

(2004). Growing intra-industry trade in manufacturing: Implications for policy. Economic & Political Weekly, 39(41), 4556–4559.

92.

Veeramani

(2007). Industry-specific determinants of intra-industry trade in India. Indian Economic Review, 42(2), 211–229.

93.

Verdoon

P. J.

(1960). The Intra-Bloc trade of BENELUX’. In Robinson

E. A. G.

(Ed.), Economic consequences of the size of nations. International Economic Association Conference Volumes. Palgrave Macmillan.

94.

Wälde

(1994). Unequal factor prices and incomplete specialization in a Heckscher-Ohlin model of endogenous growth. Journal of Economics, 59(3), 311–323.

95.

Zhang

, Witteloostuijn

, & Zhou

(2005). Chinese bilateral intra-industry trade: A panel data study for 50 countries in the 1992–2001 period. Review of World Economics, 141(3), 510–540.