To Look East or to West? An Analysis of India’s Trade Policy from Employment Perspective

Abstract

In the face of mounting protectionist sentiments, aggravated further due to COVID-19, India is redefining its trade policy. This is evident from the fact that India has decided to not join the mega-trade deal, Regional Comprehensive Economic Partnership (RCEP), giving in to the concern of massive trade deficit with China and other countries of the Association of South East Asian Nations (ASEAN). This stance is complete opposite to India’s ‘Look East Policy’. Instead, India is looking for closer trade ties with the United States (US) and the European Union (EU). In this context, this article evaluates the performance of India’s trade, in general, as well as with its Eastern (ASEAN, China, RCEP) and Western partners (US and EU), on its industrial sector employment. Employment is a better measure than trade deficit because employment gives a picture of both growth and distributional aspects of trade effects. Using a dynamic econometric model in a panel framework, it is found that, in general, while export intensity had favourable impact, import penetration had a detrimental impact on industrial sector employment during 1996–1997 to 2016–2017. Region wise analysis shows that trade with ASEAN has been employment enhancing while that with China had a negative impact on employment. On the East versus West looking trade policy debate, analysis suggests that the long run impact of exports, to both the West (EU plus USA) and the East (RCEP) partners, on employment is positive. Also, contrary to the mainstream view, the negative effects of imports, on industrial sector employment, from RCEP appears to be insignificant than that associated with imports from the EU plus USA. Thus, trade agreements with the EU and the US cannot be a substitute for dis-engaging with the resilient Eastern partners. Rather, India should strengthen its trade prospects with its East Asian partners and simultaneously look for export market in the West to enhance its employment creation.

Keywords

Trade trade policy employment KLEMS PMG estimation

JEL Classifications F13, F14, F4, J2

I. Introduction

One of the unprecedented consequences of COVID-19 pandemic is the fanning of protectionist sentiments across that world, the rising flames of which were visible in the wake of US–China trade war, attack on World Trade Organisation, etc. In such times, India’s commitment to steady and progressive trade liberalisation (since 1991) is also being tested. Recently, India pulled out of the Regional Comprehensive Economic Partnership (RCEP) trade negotiations, primarily because of India’s ballooning trade deficit with China and other members of the RCEP. India is, lately, targeting cheaper imports from China. It seems India is giving in to growing protectionist stance, which is all the more evident from its clarion call to self-reliance. The precipitating factor for such an inward-looking policy stance, especially towards the Asian countries, is the view that trade agreements, as a result of India’s Look East Policy, have benefited only the partner countries—partner countries have successfully reduced their import –export ratio, while India’s exports to these countries have remained muted (Saraswat et al., 2018). Nonetheless, India is working towards concluding free trade agreement (FTA) with the United States (US) and the European Union (EU), particularly after plugging out of the RCEP. Is India now looking ‘West’? Will this dealignment from the East be beneficial? What should be India’s trade policy, going forward, given the current global trade scenario?

In this context, the objective of this article is to examine the overall impact of trade as well as the differential impact of trade with India’s major Eastern and Western trade partners. Unlike the recent study by Saraswat et al. (2018), where the assessment of trade performance was gauged in terms of trade balance, I shall focus the research lens on employment in industrial sector. An assessment of the impact of trade on the employment in India’s industrial sector is crucial for two reasons. First, unemployment is an issue that ails Indian economy, with growth rate of employment in recent times being as low as 1 per cent (Aggarwal & Goldar, 2019). This has been exacerbated by creation of limited employment opportunities in the traditional, labour-intensive manufacturing sector (Aggarwal & Goldar, 2019). Hence, generation of employment is an important concern among policy makers.

Second, while trade deficit indicator gives an idea on the performance of trade in absolute terms, employment figures capture both the growth and distributional aspects of trade-effects. Yet, in the Indian context, literature on the effects of trade on employment is limited and inconclusive. Some studies found that trade, particularly exports, and trade reforms had a positive impact on manufacturing employment (Banga, 2005; Ghose, 2000; Goldar, 2009; Hasan et al., 2007), while other studies found import competition had detrimental effect on employment generation (Paul, 2014; Rajesh Raj & Sasidharan, 2015; Rajesh Raj & Sen, 2012) and export did not have any significant impact on employment (Rajesh Raj & Sasidharan, 2015; Sen, 2009).

Using industry-level panel data for the period from 1996–1997 to 2016–2017, the contribution of this study is to inform debate on the topical trade policy issues, namely, whether trade enhances employment, should India engage more with its Eastern trade partners (as envisaged in the ‘look East policy’) or should India now ‘look West’? Empirically, the advantage of using a panel data set up lies in the fact that dynamic aspects of trade and employment linkages as well as the short and long run impact of trade on employment can be studies.

The remainder of the article is organised as follows. In the next section, discussion on the theoretical literature on the link between trade and employment and the theoretical basis of the model employed in the study are presented. In Section 3, the sources of data used, and construction of variables are discussed. Section 4 carries out the panel diagnostic tests and elaborates on the empirical issues in estimating the model. Section 5 presents the empirical results and Section 6 concludes the discussion and provides policy implications.

II. Theoretical Underpinnings and Empirical Model

Theoretically, scale effect, composition effect and process effect (Sen, 2009) are the three channels by way which trade can impact industrial sector employment (Sen, 2009). The scale effect of trade on employment works via total output of the industrial sector. On one hand, higher exports lead to an increase in level of output and therefore, raise employment level. On the other hand, greater import penetration reduces output leading to displacement of labour (Sen, 2009).

The composition effect of trade can be explained using the standard two-factor Heckscher–Ohlin (HO) model of trade. The HO model predicts that labour-abundant developing countries will export labour-intensive goods and import relatively higher capital-intensive commodities. Thus, trade will increase the demand for labour-intensive commodities which, in turn, will lead to higher demand for labour and greater employment prospects for the trading country.

Finally, the process effect of trade is based on the Stopler–Samuelson theorem. The theorem postulates that greater trade openness leads to a change in relative factor demand which changes relative factor prices leading to factor substitution in production. Additionally, trade-induced productivity effects can also affect manufacturing employment. Trade-induced competitive pressure can force firms to use labour-saving technologies (Greenaway et al., 1999). Such productivity effects can be due to a decrease in x-inefficiency (Chand & Sen, 2002) or due to technological transfers (Jenkins & Sen, 2006).

Empirically, there are three methodological approaches to study the impact of trade on employment. These are factor content, growth accounting and labour demand approaches. The factor-content approach examines whether a change in production structure due to greater trade openness leads to an increase in the labour intensity of production and therefore, employment. To accomplish this, direct and indirect labour requirements per unit of export and import-substitutes are calculated. The factor content approach captures the composition effect of trade on employment (Rajesh Raj & Sasidharan, 2015).

The growth accounting approach measures the scale effect of trade on employment (Rajesh Raj & Sasidharan, 2015). Under this approach, changes in employment are decomposed into the effects of changes due to domestic demand, exports, imports and productivity. It is assumed that increased export orientation generates employment whereas increased import penetration reduces employment. The labour demand approach captures the process effect of trade on employment. To do this, industry-level labour demand equations are estimated, where industry-level employment is regressed against a bunch of explanatory variables (Greenaway et al., 1999; Hine & Wright, 1998).

Earlier studies on the link between trade and employment in the Indian context have mostly focused on the employment outcome of trade via productivity effects of trade (Goldar, 2009; Mitra, 2011; Malik & Mitra, 2018; Maiti, 2019; Vashisht, 2016, etc.). In this article, we are interested in impact of India’s trade on its employment in the industrial sector as measured by labour demand in the sector. To estimate the impact of trade on labour demand, following Greenway et al. (1999), we adopt a regression-based approach grounded in a dynamic model of labour demand,¹ derived from Cobb–Douglas production function:

q_{i t} = A_{i t} k_{i t}^{α} l_{i t}^{β}

(1)

where i represents industry and t represents time. In equation (1), q is real output, k is the capital stock, l is the units of labour employed in production and A is total factor productivity term and α and β are the output elasticities of capital and labour, respectively.

Assuming perfectly competitive labour markets and solving the first-order profit maximisation for labour yields

l_{i t} = \frac{β p_{i t} q_{i t}}{w_{i t}}

(2)

where p is the domestic price of the output and w is the wage earned by labour.

By substituting (1) into (2), we get

l_{i t} = β [A k_{i t}^{α} l_{i t}^{β}] / w_{i t}

(3)

and solving (3) for labour demand of industry i at time t we get:

l_{i t} = {\{\frac{p_{i t} β A k_{i t}^{α}}{w_{i t}}\}}^{(\frac{1}{1 - β})}

(4)

Equation (4) conditions labour demand on capital stock and output is allowed to vary with changes in domestic prices associated with changes in trade as this could be a crucial channel through which trade affects labour demand at industry level (Castro et al., 2007).²

It is expected that over time the total factor productivity increases with the rate of technology adoption and increases in x-efficiency (Chand & Sen, 2002). Now, both technological adoption and reduction of x-inefficiency are correlated to trade openness, therefore $A_{i t} p_{i t}$ is assumed to be function of import penetration and export penetration (or export demand):

A p_{i t} = e^{(γ_{0} T_{i})} M_{i t}^{(γ_{1} + (\frac{1}{τ^{M}}))} X_{i t}^{(γ_{2} + (\frac{1}{τ^{X}}))}

(5)

where $T_{i}$ is time trend, M_it and X_it are import and export intensity, respectively. $τ^{M}$ is import elasticity of demand whereas $τ^{X}$ is export elasticity of demand. $γ_{0}$ measures the time trend in total factor productivity (TFP). $γ_{1}$ and $γ_{2}$ are the elasticities of TFP w.r.t. imports and exports, respectively.

Substituting (5) into (4) and taking logs we get following equation which forms the basis for our empirical specification:

ln (l_{i t}) = α_{0} + α_{i} ln (K_{i t}) + α_{2} ln (w_{i t}) + α_{3} ln (M_{i t}) + α_{4} ln (X_{i t}) + α_{5} T + α_{6} I + ε

(6)

Augmented versions of this model, adding the trade shares of India’s trade partners in Equation (6), will also be estimated to assess the marginal impact of imports from and exports to the partners.

III. Data Sources and Construction of Variables

Preparing the Database for our study

Given the empirical specification, our study requires industry wise annual statistics on India’s trade (exports and imports) with its trading partners in industrial goods, the measure of industrial output, that is, value added, and on employment, capital stock and wages for India’s industrial sector. The period of our study is 1996–1997 to 2016–2017. The data on India’s bilateral trade with its partners is sourced from the Trade Statistics of Ministry of Commerce and Industry (MoCI) which compiles and publishes the export and import data on merchandise goods for registered manufacturing sectors. I have collected the data on value added and capital stock at constant 2011–2012 prices (in Rs. crore) and employment (in 1000s) from India KLEMS database, the latest version of which is available as of 2019.³ The source of data on wages is the Annual Survey of Industries (ASI), taken from Centre for Monitoring Indian Economy Pvt Ltd. database. As the industrial classification codes for the data pertaining to the variables of interest differ, first, I have to do an extensive mapping exercise so as to make all the data compatible and classified according to a single nomenclature. In brief, I want to map merchandise exports (which are goods, coded as per Harmonised System (HS)) to industries (which are classified by National Industrial Classification (NIC) code in India). Finally, it is necessary to make the data defined in accordance with India KLEMS industry codes for carrying out regression analysis. The step by step process of mapping is explained in the following paragraphs.

First, the ASI classifies manufacturing activities according to the NIC schedule, which in turn is based on International Standard Industrial Classification (ISIC) nomenclature of the United Nations. The period of study is from 1996–97 to 2016–17, during which the NIC nomenclature has been restructured three times, namely NIC-1998, NIC-2004 and NIC-2008 which follow the ISIC rev 3, ISIC rev3.1 and ISIC rev4, respectively. Using the three concordance tables provided by the ASI, namely, NIC-1998, NIC-2004 and NIC-2008,⁴ and the concordance table between ISIC rev 3 and ISIC rev 4⁵ provided by United Nations Statistical Division, a consistent time series of 15 industrial sector at 2-digit level corresponding to ISIC rev 3 covering broad industry groups is built.

Second, the MoCI adopts the Harmonised Commodity Description and Coding System (HS) for commodity classification and provides trade statistics according to HS nomenclature. As trade and industry statistics are reported according to different nomenclatures, namely, HS code and ISIC rev 3, respectively, industry-trade concordance exercise is also required for the study. To do this, trade data with the HS code nomenclature is mapped with to ISIC rev 3 codes using a concordance table accessed from World Integrated Trade Solution,⁶ using minimum discretion where perfect match was not available.

Lastly, in order to make industrial classification of all the variables aligned to India’s KLEMS code, I first use the concordance table between NIC code and KLEMS industry code, available in India’s KLEMS manual, and convert the NIC codes (which are now in ISIC rev 3 codes) into KLEMS industry code with minimum discretion. Next, the HS codes (which are now in ISIC rev 3 codes) are made compatible with India’s KLEMS industry nomenclature, using minimum discretion.⁷ Thus, the industry classification codes for the entire dataset is now in accordance with KLEMS.

To get a bit more insight into the nature of the industries, I have re-organised and grouped the 15 industries, classified according to the KLEMS industry code, into labour-intensive and non-labour-intensive industries. For dividing the industries into labour- and non-labour-intensive industries, I start with the basic idea that an industry that requires a large amount of labour as compared to capital for production is a labour-intensive industry, that is, if the labour cost outweighs the capital cost, the production process is said to be labour-intensive as the major cost incurred by the industry is cost of paying the workers. Using this premise, I calculate the following ratio for each industry:

R_{i} = \frac{\{[\frac{L_{i}}{K_{i}}] - \min_{i = 1 t o n} [\frac{L_{i}}{K_{i}}]\}}{\{\max_{i = 1 t o n} [\frac{L_{i}}{K_{i}}]\} - \{\min_{i = 1 t o n} [\frac{L_{i}}{K_{i}}]\}}

(7)

R_{i}

will lie between 0 and 1; with

R_{i} = 1

representing the most labour-intensive industry and

R_{i} = 0

representing the least labour-intensive industry. We rank the industries using this ratio and take

R_{i} = 0.5

as the cut-off point to classify industries having ratio more than 0.5 to labour-intensive. The list of industries that has been used in our study is given in the following Table 1.

Table 1.

List of Industries Included in the Sample

Sr. No	Name of Industry	Labour/Non-Labour-Intensive
1	Agriculture, Hunting, Forestry and Fishing	Labour intensive
2	Mining and Quarrying	Non-Labour intensive
3	Food Products, Beverages and Tobacco	Labour intensive
4	Textiles, Textile Products, Leather and Footwear	Labour intensive
5	Wood and Products of wood	Labour intensive
6	Pulp, Paper, Paper products, Printing and Publishing	Labour intensive
7	Coke, Refined Petroleum Products and Nuclear fuel	Non-Labour intensive
8	Chemicals and Chemical Products	Non-Labour intensive
9	Rubber and Plastic Products	Non-Labour intensive
10	Other Non-Metallic Mineral Products	Labour intensive
11	Basic Metals and Fabricated Metal Products	Non-Labour intensive
12	Machinery, nec.	Labour intensive
13	Electrical and Optical Equipment	Labour intensive
14	Transport Equipment	Non-Labour intensive
15	Manufacturing, nec; recycling	Labour intensive

Source: Author’s classification using KLEMS database.

Construction of Trade related Indices

To study the impact of trade on India’s industrial sector employment, we constructed trade related variables such as import penetration index and export intensity for a panel of 15 industrial sectors for the period 1996–1997 to 2016–2017. We use trade indices rather than absolute values of exports and imports because export intensity and import penetration indices measure the performance and importance of exports and imports respectively.

The import penetration index (IMP) for industrial sector i in year, t is defined as the ratio of total imports to domestic demand (i.e., the difference between output and exports):

IMP = (\frac{M_{i t}}{Q_{i t} - X_{i t} + M_{i t}})

(8)

where M represents value imports, X represents value exports and Q represents output or the total value added.

The export intensity (EXI) is defined as the ratio of total exports for sector i in year t to total production of that sector in year t:

EXI = (\frac{X_{i t}}{Q_{i t}})

(9)

The summary statistics of the variables are provided in Table 2.

Table 2.

Summary Statistics of Variables

Variables	N	Mean	Std	Min	Max
Ln(Employment)	313	8.010	1.658	4.267	12.46
Ln(Capital)	313	12.44	0.985	9.893	15.13
Ln(Wage)	313	10.63	1.883	4.002	13.75
Ln(Import Penetration)	313	–3.492	1.561	–7.981	– 0.305
Ln(Export Intensity)	313	– 4.285	1.943	–12.62	– 0.269

Source: Author’s Analysis.

IV. Panel Diagnostic Tests and Empirical Strategy

The occurrence of cross-section dependence is common in macro panels, with higher time dimension, (Pesaran & Smith, 1995). Cross-section dependence (CSD) may arise when the cross-sectional dimensions, that is, industries react in the same manner to a given shock. To assess the presence of CSD, Pesaran (2004) cross-sectional dependence (CD) test is applied and the result is reported in Table 3. The result shows a failure to reject the null of cross independence.

Since cross-dependence is not an issue in the panel, I proceed with first generational panel unit root test, namely, Fisher ADF unit root test and the Im et al. (2003) unit root test to examine the stationarity properties of interest variables. The results are reported in Table 4.

Table 3.

Results of Cross-Sectional Dependence Test

Test	Statistic	Probability
Pesaran’s test of cross-sectional independence	–1.236	1.7837

Source: Author’s Analysis.

Note: Null hypothesis: cross-sectional independence (CD ~ (0.1).

Table 4.

Unit Root Results

	ADF-Fisher			Im, Pesaran Shin
Variables	Level	1st Difference	Integration	Level	1st Difference	Integration
ln(Employment)	25.184	77.929***	I(1)	6.059	–3.271***	I(1)
ln(Capital)	128.859***		I(0)	1.766	–4.598***	I(1)
ln(Wage)	1.420	316.926***	I(1)	10.713	–8.341***	I(1)
ln(Import Penetration)	20.221	272.428***	I(1)	0.669	–8.767***	I(1)
ln(Share of Imports from ASEAN)	52.809**		I(0)	–1.408*		I(0)
ln(Share of Imports from China)	100.241***		I(0)	–3.296***		I(0)
ln(Share of Imports from USA)	62.231***		I(0)	–1.487*		I(0)
ln(Share of Imports from EU)	50.291**		I(0)	0.514	–9.962***	I(1)
ln(Share of Imports from RCEP)	52.500**		I(0)	–1.500*		I(0)
ln(Share of Imports from USA+ EU)	46.109**		I(0)	0.506	–9.400***	I(1)
ln(Export Intensity)	11.732	267.594***	I(1)	1.659	–8.549***	I(1)
ln(Share of Exports to ASEAN)	70.180***		I(0)	–3.546***		I(0)
ln(Share of Exports to China)	77.482***		I(0)	–3.516***		I(0)
ln(Share of Exports to USA)	42.733*		I(0)	–1.487*		I(0)
ln(Share of Exports from EU)	51.892**		I(0)	–1.529*		I(0)
ln(Share of Exports to RCEP)	61.6416***		I(0)	–3.093***		I(0)
ln(Share of Exports to USA + EU)	41.2512*		I(0)	–1.547*		I(0)

Source: Author’s analysis.

Notes: (i) Null hypothesis is the presence of a unit root. (ii) *** p < .01, ** p < .05, * p < .1

The results reveal that all the variables are integrated of mixed order at either levels or first difference. Subsequent to unit root tests, I investigate the long-run cointegrating relationships between variables. The Pedroni cointegration test advanced by Pedroni (1999) in conjunction with the Kao (1999) cointegration test is employed to investigate the existence of convergence among the variables presented in Tables 5 and 6. Kao is an improvement over Pedroni as it can take into account variables more than seven in number.

Table 5.

Pedroni’s (1999) Residual Cointegration Test Results.

	t-statistic	p-value
Model 1
Modified Phillips–Perron	2.298	.011
Phillips–Perron	–2.046	.020
Augmented Dickey-Fuller	–.290	.386
Model 2 (with ASEAN Trade Share)
Modified Phillips–Perron	3.995	.000
Phillips–Perron	–1.586	.056
Augmented Dickey–Fuller	.265	.395
Model 3 (with China Trade Share)
Modified Phillips–Perron	3.631	.000
Phillips–Perron	–2.351	.009
Augmented Dickey–Fuller	–1.078	.141
Model 4 (with USA Trade Share)
Modified Phillips–Perron	4.212	.000
Phillips–Perron	–0.931	.176
Augmented Dickey–Fuller	–0.815	.208
Model 5 (with EU Trade Share)
Modified Phillips–Perron	3.617	.000
Phillips–Perron	–2.710	.003
Augmented Dickey–Fuller	–1.208	.114

Source: Author’s analysis.

Note: The null hypothesis is that the variables are not cointegrated. Lag selection is based on Schwarz Information Criterion.

Table 6.

Kao’s (1999) Residual Cointegration Test Results.

	t-statistic	p-value
ADF Model 1	4.427	.000
ADF Model 2 (with ASEAN Trade Share)	4.654	.000
ADF Model 3 (with China Trade Share)	3.933	.000
ADF Model 4 (with USA Trade Share)	3.8485	.000
ADF Model 5 (with EU Trade Share)	3.6056	.000
ADF Model 6 (with RCEP and EU+USA Trade Shares)	4.2585	.000

Source: Author’s analysis.

Note: The null hypothesis is that the variables are not cointegrated. Lag selection is based on Schwarz Information Criterion.

Both the cointegration tests rejects the null hypothesis of no cointegration, thus supporting the presence of long-run equilibrium relationship between the variables. Now, in order to estimate the magnitude of cointegration in terms of coefficients, panel autoregressive distributed lag (ARDL) modelling technique that allows for the presence of mixed I(0) and I(1) variables needs to be employed.

Therefore, I adopt an estimation approach of Pooled Mean Group (PMG), which is a panel ARDL technique proposed by Pesaran and Smith (1995, 1999), to explore the short run and long run dynamics between the dependent variables and its explanatory variables. The econometric equation is defined as follows:

\begin{array}{l} Δ ln Employment_{i, t} = \\ \sum_{j = 1}^{p - 1} γ_{j}^{i} Δ ln Employment_{i, t - j} + \sum_{j = 0}^{q - 1} δ_{j}^{i} Δ X_{i, t - j} + φ^{i} [ln Employment_{i, t - 1} - \{β_{0}^{i} + β_{1}^{i} X_{i, t - j}\}] + \in_{i t} \end{array}

(10)

where i denote industries, t stands for time period, p represents lag length of dependent variable, q denotes the lag length of independent variables, ln Employment is the natural logarithm of industrial sector employment, X is the vector of explanatory variables such as ln(Capital), ln(Wage), import penetration and export penetration. γ denotes lag coefficient of Employment variable, δ represents the short-run coefficients of explanatory variables which vary across industries, φ is the coefficient of speed of adjustment to the long-run equilibrium, and β denotes the long-run coefficients that are homogenous across industries. These coefficients are derived from Equation (11).

ln Employment_{i, t} = β_{0}^{i} + β_{1}^{i} X_{i, t} + μ_{i, t} where μ_{i, t} ~ I (0)

(11)

The PMG-ARDL technique has a number of econometric advantages relative to other traditional panel estimation techniques. First, it accommodates the endogeneity issues in econometric modelling. It will control for potential endogeneity, in this case, arising out of the fact that the labour employment is likely to show inertia of variables, by introducing optimal lag length of the variables. The PMG-ARDL technique also remains robust to lag orders and outliers (Pesaran et al., 1999).Second, it is appropriate for heterogenous dynamic industry-specific panel data where long-run is characterised by conditions expected to be homogenous across industries while the short-run adjustment depends on industry-specific factors such as wage-price flexibility, vulnerability to domestic and external shocks and government policies. It estimates both long- and short-run coefficients simultaneously. Finally, as discussed earlier, PMG-ARDL technique is flexible in terms of its application in case of mixed order of cointegration—be it I(0) or/ and I(1). Since an appropriate ARDL lag-structure should be determined by some consistent information criterion, we impose the lag structure (1,1,1,1,1) for employment, capital stock, wages, export intensity and import penetration, respectively, which is based on the Schwartz Bayesian criterion.

V. Results and Inferences

Table 7 reports the results of PMG estimations for the base model. In Table 7, the independent variables converge to their long-run path by a magnitude −0.064, which is statistically significant at 1 per cent level, thereby, affirming the equilibrium relationship between the variables. The results indicate that while import penetration has a negative and significant impact on industrial sector employment in India in the long run, export intensity has a positive and significant impact on employment in the long run.⁸ In the short run, both import penetration and export intensity are negatively related to employment. This finding corroborates with Dutt et al. (2009) who found that unemployment rose in the short run as a response to trade liberalisation. In the long run, the authors found a reversal of that rise and an ensuing decline in unemployment. However, in the short run, neither import penetration nor export intensity has any significant impact on employment.⁹ The economic intuition for getting insignificant short run coefficient is that the impact of any economic shock, such as trade, can only be significantly observed with lag, that is, in the long run.

The long run panel fitted model also shows that capital significantly exerts a positive impact on employment but a negative impact in the short run. A plausible explanation is that an increase in physical capital stock had a complementary effect on employment and has contributed to the increase in industrial employment in India. Also, the impact of wage on employment is significant in both long and short run—its impact is negative in the long run and positive in the short run. Intuitively, a rise in wage rate increases the cost of hiring labour. Therefore, firms switch to employing labour-saving technology leading to a decline in industrial employment.

Table 7.

Results of PMG-ARDL

Dependent Variable: ln(Employment)
Variable	Coefficient	Std. Error	p-value
Long run
ln(Capital)	0.506***	0.084	.000
ln(Wage)	–0.438***	0.018	.000
ln(Import Penetration)	–0.095**	0.038	.013
ln(Export Intensity)	0.055**	0.025	.031

Short run
Error Correction Term	–0.064***	0.024	.009
∆ ln(Capital)	–0.085**	0.040	.033
∆ ln(Wage)	0.069**	0.035	.046
∆ ln(Import Penetration)	–0.007	0.010	.469
∆ ln(Export Intensity)	–0.003	0.016	.837
Intercept	0.418**	0.160	.009

Source: Author’s analysis.

Notes: The fitted model is based on maximum lag 1 as suggested by Akaike information criterion with 298 observations

*** p < .01, ** p < .05, * p < .1

To assess the marginal impact of imports and exports with major trading partners of India, Table 8 reports PMG estimation results, including the trade shares partners, namely, ASEAN, China, US and EU. India has a decade-old FTA with ASEAN and China is India’s second largest trade partner after USA. Apart from this, EU is also India’s one of the largest and important trading partner. The estimated long run coefficients of capital stock and wage rate have the expected signs and are significant and stable across all specifications; except for in the case of specification with ASEAN where capital stock is insignificant.

Table 8.

Results of PMG–ARDL, Including Trade Shares by Partners

Dependent Variable: ln(Employment)
	ASEAN		China		USA			EU
Variables	Coefficient	Std. Error	Coefficient	Std. Error	Coefficient	Std. Error		Coefficient	Std. Error
Long run
ln(Capital)	0.098	0.065	0.566***	0.135	0.694***		0.216	1.009***	0.059
ln(Wage)	–0.372***	0.039	–0.406***	0.028	–0.695***		0.144	–0.355***	0.037
ln(Import Penetration)	–0.294***	0.058	–0.184***	0.060	–0.248**		0.103	–0.399***	0.038
ln(Share of Imports from ASEAN)	0.204***	0.026
ln(Share of Imports from China)			–0.038**	0.031
ln(Share of Imports from USA)					0.077		0.062
ln(Share of Imports from EU)								–0.017**	0.017
ln(Export Intensity)	0.289***	0.035	0.108***	0.040	0.143**		0.063	0.228***	0.037
ln(Share of Exports to ASEAN)	0.082***	0.022
ln(Share of Exports to China)			–0.024*	0.015
ln(Share of Exports to USA)					0.353**		0.132
ln(Share of Exports from EU)								0.043	0.047
Short run
Error Correction Term	–0.061**	0.026	–0.053***	0.020	–0.033**		0.014	–0.043*	0.034
∆ ln(Capital)	–0.056	0.059	–0.136**	0.058	–0.072		0.064	–0.078	0.066
∆ ln(Wage)	0.082*	0.045	0.083***	0.031	0.064*		0.034	0.055*	0.033
∆ ln(Import Penetration)	0.000	0.008	0.002	0.011	0.002		0.010	–0.005	0.011
∆ ln(Share of Imports from ASEAN)	–0.005	0.011
∆ ln(Share of Imports from China)			0.021	0.022
∆ ln(Share of Imports from USA)					–0.006		0.008
∆ ln(Share of Imports from EU)								–0.001	0.013
∆ ln(Export Intensity)	–0.006	0.020	–0.002	0.013	–0.017		0.018	0.003	0.014
∆ ln(Share of Exports to ASEAN)	0.004	0.009
∆ ln(Share of Exports to China)			0.004	0.006
∆ ln(Share of Exports to USA)					–0.002		0.020
∆ ln(Share of Exports from EU)								–0.016	0.022
Intercept	0.650**	0.235	0.259***	0.097	0.251**		0.105	–0.102**	0.052

Source: Author’s analysis.

Notes: The fitted model is based on maximum lag 1 as suggested by Akaike information criterion with 298 observations

*** p < .01, ** p < .05, * p < .1

The marginal impact of imports from and exports to ASEAN on employment is positive and significant in the long run and is insignificant in the short run. The positive effect of import share could be a manifestation of upstream import penetration, mainly in the food products, beverages and tobacco industry which is a labour-intensive industry (see Appendix A2). However, an examination of industry-wise trade with ASEAN reveals that the growth in share of exports in the food products industry, like other industries, has been slow as compared to imports in this industry. Thus, the potential of employment generation via exports has not been harnessed full.

In the case of China, the estimated coefficient of import share is negative and significant in the long run. This is not surprising as India’s imports from China is mostly concentrated in labour-intensive industries such textiles, manufacture of non-metallic mineral products, machinery and equipment, and electrical and optical equipment (see Appendix A3).¹⁰ Thus, import competition from China has had a detrimental effect on India’s industrial sector employment. What is all the more surprising, however, is the estimated coefficient of export share is also negative and significant in the long run. The results show that a rise of share of exports to China by 1 per cent causes around 0.024 per cent decline in employment. This indicates that there could be trade induced preference for skilled labour in industries (mainly agricultural industry) exporting to China, thus displacing unskilled labour.

As far as trade with USA is concerned, imports from USA do not seem to have any significant impact on the employment. This insignificant result may be because imports from USA has been increasingly in agricultural items, chemical and chemical products and rubber and plastic product, among other non-labour-intensive industries, which can hardly have any disadvantageous impact on employment. On the other hand, share of exports to USA has had a positive impact on employment. The explanation to this finding is quite straightforward – USA is India’s export market in labour-intensive industries such as electrical machinery, medical and optical equipment, mechanical appliances as well as textiles and manufacturing of wood and wood products (see Appendix A3).

Similar to the results for China, the long run impact of import from EU is negative and significant. However, the share of export to EU hardly has any impact on the employment. This is because of the fact that the India’s importance in EU’s imports has not been significant. According to the latest data, India is tenth largest partner for EU imports and the share of EU’s import from India is only about 2 per cent of its total imports.¹¹ The negative impact could be because of the import competition faced by labour-intensive industries such as manufacturing of paper and paper products, non-metallic mineral products, machinery and equipment and electrical and optical equipment (see Appendix A4), which may have led to the decline in output thereby leading to decline in demand for labour these industries. The insignificant impact of India’s exports to EU

Apropos to the policy debate on whether India should open its market, going forward, to its Eastern partners or the Western partners, an analysis of impact of India’s trade with major Eastern and Western partners on employment is carried out. Table 9 reports the findings of the analysis of India’s trade with RCEP (East) versus EU plus USA (West). We have taken RCEP because India was in negotiation with this group for the last few years before plugging out of RCEP in December 2019. In the West, USA and the EU are the most important trading partners of India and India is currently negotiating trade deals with the EU and the USA.

As can be inferred from Table 9, the long run impact of exports, to both the West and the East partners, on India’s industrial employment is positive and significant. Nonetheless, the magnitude of impact seems to be higher for RCEP as compared to EU and USA combined. The estimation result also shows that the import shares have negative impact on employment. However, the impact is not significant in the case of RCEP.

Table 9.

Results of PMG–ARDL, Including Trade Shares by East and West Partners

Dependent Variable: ln(Employment)
Variables	Coefficient	Std. Error
Long run
ln(Capital)	1.145***	0.129
ln(Wage)	-0.433***	0.070
ln(Import Penetration)	-0.456***	0.070
ln(Share of Imports from RCEP)	-0.002	0.105
ln(Share of Imports from USA+ EU)	-0.193*	0.103
ln(Export Intensity)	0.179***	0.064
ln(Share of Exports to RCEP)	0.285***	0.068
ln(Share of Exports to USA + EU)	0.106**	0.093
Short run
Error Correction Term	-0.044*	0.035
∆ln(Capital)	-0.052	0.077
∆ln(Wage)	0.051	0.028
∆ln(Import Penetration)	-0.006*	0.020
∆ln(Share of Imports from ASEAN)	0.043	0.037
∆ln(Share of Imports from USA + EU)	0.011	0.037
∆ ln(Export Intensity)	0.016	0.010
∆ln(Share of Exports to RCEP)	-0.001	0.028
∆ln(Share of Exports to USA+EU)	-0.032	0.038
Intercept	-0.125	0.097

Source: Author’s analysis.

Notes: The fitted model is based on maximum lag 1 as suggested by Akaike information criterion with 298 observations.

*** p < .01, ** p < .05, * p < .1

Furthermore, it is observed that the inclusion of trade shares of India’s East and West partners does not lead to any change in the results related to wage rate and physical capital stock. Akin to the PMG results reported in Tables 7 and 8, an increase in wage rate is associated with a decline in industrial-level employment whereas an increase in physical capital stock is associated with an increase in industrial-level employment. Again, in line with results reported in Tables 7 and 8, while the estimated coefficient for import penetration negative and significant, that for export intensity is positive and significant in the long run. Also, the independent variables converge to their long-run path.

VI. Concluding Remarks and Policy Outlook

This study investigates the impact of overall trade as well as the effect of India’s trade with its major Eastern and Western partners on employment outcome in the India’s industrial sector. With regard to overall trade, the results reveal that while import penetration has a negative and significant impact, export intensity has a positive and significant impact on industrial sector employment in India in the long run. This finding is in line with earlier studies on this issue, namely, Banga (2005), Goldar (2009) and Paul (2014). The analysis of the differential effect of India’s trade with ASEAN, China, EU and the USA provides some interesting results. The long run impact of India’s trade with the ASEAN has been employment enhancing mainly because of the manifestation of upstream import penetration in the food products, beverages and tobacco industry which is a labour-intensive industry. However, the growth in share of exports in this industry, like other industries, has been slow as compared to imports in this industry. Thus, the potential of employment generation via exports has not been harnessed fully.

Both import competition from and share of exports to China have had a detrimental effect on India’s industrial sector employment. This could be a reflection of trade induced preference for skilled labour in industries exporting to China, thus dislocating unskilled labour. Similarly, share of imports from the EU too had a negative impact on the employment primarily because of the competition faced by labour-intensive industries. Nevertheless, share of exports to EU had insignificant impact on employment. As far as trade with USA is concerned, imports from USA do not seem to have any significant impact on the employment as they are mostly in non-labour-intensive industries. On the contrary, share of exports to USA has had a positive impact on employment as USA is India’s export market in labour-intensive industries.

On the East versus West looking trade policy debate, analysis suggests that the long run impact of exports, to both the West (EU plus USA) and the East (RCEP) partners, on India’s industrial employment is positive and significant, with the impact imposed by exports to RCEP has higher magnitude as compared to that imposed by exports to EU and USA combined. Also, contrary to the mainstream view, the negative effects of imports from RCEP appears to be less noticeable than that associated with imports from the EU plus USA.

The findings of this study have some policy implications. First, as opposed to protectionist stance, India has to ramp up trade engagement and not shy away from it. In general, exports have the potential to create employment opportunities. That said, India can reap benefits form trade agreements only if builds capacity in terms of producing goods at competitive prices for the export market. But this does not mean that exports should be considered in isolation and export promotion should be seen as sole policy objective. Instead, policy outlook should aim at promoting both exports and imports. In addition, structural policy reforms in labour-intensive industries in terms of removing supply side bottlenecks, namely, infrastructure, credit, energy deficit, inferior logistics, rigid land and labour laws are needed.

Second, as far as engagement with East and West is concerned, it is in India’s interest to pursue look East policy as trade partnership with countries having common border, culture and ethnicity (Anderson & Van Wincoop, 2004; Guiso et al., 2004) is likely to be more successful. In fact, our results support this fact from employment point of view. Eastern Asia is resilient regional value chain hub. Strategically, India should consider deeper integration with ASEAN and import intermediate and capital goods at internationally competitive prices. This, in turn, will improve productive capacity of India’s industrial sector and can make Indian exports competitive, thereby generating employment opportunities in the industrial sector. India is also looking to build closer trade ties with the West, especially after pulling out of RCEP trade deal owing to huge trade deficit with China and other members of RCEP. However, it would be very difficult for India to get trade agreement with EU and USA exclusively on its terms. Trade agreements with the EU and the US cannot be a substitute for dis-engaging with the resilient Eastern partners. Hence, India should reconsider RCEP option and strengthen its trade prospects with the ASEAN, even while negotiating FTA with the US and the EU. As our results suggest, to make trade beneficial for employment, India should strengthen intra-industry trade with its East Asian partners and simultaneously look for export market in the West. Looking West can never be complete without strengthening India’s Eastern trade ties.

Footnotes

Declaration of Conflicting Interests

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

Funding

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

Notes

ORCID iD

Devasmita Jena

Appendix A

A1: The following graph presents the orthogonal impulse response functions (IRFs). The IRFs are obtained from panel VAR estimation of the variables pertaining to employment, export intensity and import penetration.

Appendix A5. India’s Trade with EU by Category

	Percentage Share in Total Imports				Percentage Share in Total Exports
Industries	1996–97	2003–04	2010–11	2016–17	1996–97	2003–04	2010–11	2016–17
Agriculture, hunting, forestry and fishing	2	4	1	5	21	20	18	15
Mining and quarrying	11	7	2	3	15	10	4	7
Food products, beverages and tobacco	4	3	4	4	22	19	15	13
Textiles, textile products, leather and footwear	18	14	11	10	40	37	34	29
Wood and products of wood	4	4	6	8	16	26	26	20
pulp, paper, paper products, printing and publishing	34	30	28	24	13	15	13	12
coke, refined petroleum products and nuclear fuel	1	0	0	1	0	11	20	7
Chemicals and chemical products	22	21	14	14	25	22	20	18
Rubber and plastic products	32	31	19	18	26	20	21	20
other non-metallic mineral products	46	36	23	17	21	26	26	20
Basic metals and fabricated metal products	51	36	14	12	18	14	11	13
Machinery, nec.	49	33	31	25	19	22	22	21
electrical and optical equipment	36	27	21	13	23	28	29	24
Transport equipment	25	22	31	32	27	34	19	19
Manufacturing, nec; recycling	20	19	10	4	35	26	6	27

Source: Author’s calculations using KLEMS and MoCI database.

References

Aggarwal

S. C.

Goldar

(2019). Structure and growth of employment: Evidence from India KLEMS data. India Growth and Development Review . 12(2), https://doi.org/10.1108/IGDR-10-2018-0103.

Anderson

J. E.

Van Wincoop

(2004). Trade costs. Journal of Economic Literature , 42(September), 691–751.

Banga

(2005). Impact of liberalization on wages and employment in Indian manufacturing industries (WP No. 153). Indian Council for Research on International Economic Relations, February.

Castro

Olarreaga

Saslavsky

(2007). The impact of trade with China and India on Argentina’s manufacturing employment . (World Bank Policy Research Working Paper No. 4153).

Chand

Sen

(2002). Trade liberalization and productivity growth: Evidence from Indian manufacturing. Review of Development Economics , 6(1), 120–132.

Dutt

Mitra

Ranjan

(2009). International trade and unemployment: Theory and cross-national evidence. Journal of International Economics , 78(1), 32–44.

Goldar

(2009). Impact of trade on employment generation in manufacturing in India (Working Paper Series No. E/297/2009). Institute of Economic Growth, University of Delhi.

Ghose

K. A.

(2000). Trade liberalization and manufacturing employment . (Employment Paper 2000/3). International Labour Office Geneva.

Greenaway

Robert

Wright

(1999). An empirical assessment of impact of trade on employment in the United Kingdom. European Journal of Political Economy , 15(3), 485–500.

10.

Guiso

Sapienza

Zingales

(2004). Cultural biases in economic exchange (NBER Working Paper, 11005). Cambridge.

11.

Hasan

Mitra

Ramaswamy

K. V.

(2007). Trade reforms, labor regulations, and labor-demand elasticities: Empirical evidence from India. The Review of Economics and Statistics , 89(3), 466–481.

12.

Hine

R. C.

Wright

P. W.

(1998). Trade with low wage economies, employment and productivity in UK manufacturing. The Economic Journal , 108(450), 1500–1510.

13.

K. S.

Pesaran

Shin

(2003). Testing for unit roots in heterogeneous panels. Journal of Econometrics 115(1), 53–74.

14.

Jenkins

R. O.

Sen

(2006). International trade and manufacturing employment in the south: Four country case-studies. Oxford Development Studies , 34(3), 299–322.

15.

Kao

(1999). Spurious regression and residual-based tests for cointegration in panel data. Journal of Econometrics 90(1), 1–44.

16.

Krishna

K. L.

Goldar

Das

D. K.

Aggarwal

S. C.

Erumban

A. A.

Das

P. C.

(2020). Manufacturing productivity in India: The role of foreign sourcing of inputs and domestic capacity building. In Barbara

M. Fraumeni.

(Ed.), Measuring Economic Growth and Productivity: Foundations, KLEMS Production Models, and Extensions , 2020. https://doi.org/10.1016/B978-0-12-817596-5.00006-8.

17.

Maiti

(2019). Trade, labor share, and productivity in India’s industries (ADBI Working Paper, No. 926). Asia Development Bank Institute.

18.

Malik

Mitra

(2018). Technology, TFPG and employment: A panel data analysis (NILERD Working Paper Series 3/2018).

19.

Mitra

(2011). Trade in services: Impact on employment in India. The Social Science Journal 48(2011), 72–93.

20.

Paul

(2014). Import intensity and its impact on exports, output and employment (Working Paper No. 167). ISID.

21.

Pedroni

(1999). Critical values for cointegration tests in heterogeneous panels with multiple regressors. Oxford Bulletin of Economics and Statistics 61(Special Issue), 653–670.

22.

Pesaran

(2004). General diagnostic tests for cross section dependence in panels (Cambridge Working Papers in Economics No. 0435). Faculty of Economics, University of Cambridge.

23.

Pesaran

M. H.

Smith

(1995). Estimating long-run relationships from dynamic heterogeneous panels. Journal of Econometrics , 68(1), 79–113.

24.

Pesaran

M. H.

Shin

Smith

R. P.

(1999). Pooled mean group estimation of dynamic heterogeneous panels. Journal of the American Statistical Association , 94(446), 621–634.

25.

Pesaran

M. H.

Smith

(1995). Estimating long-run relationships from dynamic heterogeneous panels. Journal of Econometrics , 68(1), 79–113.

26.

Rajesh Raj

S. N.

Sasidharan

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

27.

Rajesh Raj

Sen

(2012). Did international trade destroy or create jobs in Indian manufacturing? European Journal of Development Research 24, 359–381. https://doi.org/10.1057/ejdr.2011.19

28.

Saraswat

V. K.

Priya

Ghosh

(2018). A note on free trade agreements and their costs. NITI Aayog . https://www.niti.gov.in/writereaddata/files/document_publication/FTA-NITI-FINAL.pdf

29.

Sen

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

30.

Vashisht

(2016). Creating manufacturing jobs in India: Has openness to trade really helped. Journal of Asian Economics , 42, 53–64.