Impact of global value chains on total factor productivity: The case of Indian manufacturing

• Published date: 01 May 2022
• Author: Karishma Banga
• Date: 01 May 2022
• DOI: http://doi.org/10.1111/rode.12867

704 

|

     wileyonlinelibrary.com/journal/rode  Rev Dev Econ. 2022;26:704–735.

© 2022 John Wiley & Sons Ltd

Received: 23 June 2021 

|

  Revised: 18 January 2022 

|

  Accepted: 18 January 2022

DOI: 10.1111/rode.12867  

REGULAR ARTICLE

Impact of global value chains on total factor

productivity: The case of Indian manufacturing

KarishmaBanga

Digital and Technology Cluster, Institute 

of Development Studies, Brighton, 

United Kingdom

Correspondence

Karishma Banga, Digital and Technology 

Cluster, Institute of Development 

Studies, Library Road, Falmer, Brighton 

BN1 9RE, United Kingdom.

Email: k.banga@ids.ac.uk

Abstract

With  production  increasingly  fragmented  across  bor-

ders,  global  value  chains  (GVCs)  form  a  key  feature 

of  the  world  economy.  A  rising  body  of  literature  has 

documented productivity gains from linking into GVCs 

for developed- country firms, but studies on the same for 

developing countries are scant. In this paper, we empiri-

cally examine  whether GVCs  can increase  total factor 

productivity  (TFP)  in  developing  countries,  using  an 

unbalanced panel of Indian manufacturing firms in the 

period 2000/2001– 2014/2015 from the Prowess data set. 

To address econometric  issues of self- selection  and en-

dogeneity, we use a two- stage empirical strategy of  pro-

pensity score matching and system generalized method 

of moments. The results indicate that linking into GVCs 

has a positive and significant impact  on firm TFP, even 

after  accounting  for  self- selection  of  more  productive 

firms into GVCs.  Once linked, increasing GVC  embed-

dedness, captured through firm- level vertical specializa-

tion, also  has a positive and  significant impact on firm 

productivity.  But  productivity  gains  are  not  homog-

enous; GVC  firms that  link into relational  governance 

chains,  captured  through  the  share  of  skilled  workers 

and  infrastructure  and  communication  index  relative 

to industry,  have 20% higher  productivity growth than 

firms in other chains. Older and more liquid GVC firms 

are also more productive. Designing policies to facilitate 

GVC  linkages  and  firm  technological  capabilities  can 

thus foster productivity gains in developing countries.

|

705

BANGA

1

|

INTRODUCTION

A large body of literature has examined global value chains (GVCs) as a driver of manufacturing 

productivity growth, which  can in turn  lead to creation of new jobs,  reduction in poverty, and 

economic development.  A range of  channels have  been identified  through which linking  into 

GVCs can  increase firm  productivity.  Recent theoretical models  have shown  that productivity 

gains associated  with GVCs may accrue  from countries participating  in the finer  international 

division  of  labor, which  can  lead  to  factor- augmenting  technical  change  (Grossman  &  Rossi- 

Hansberg,2008). Firms  can focus on specialized tasks,  in which they enjoy core competencies, 

while offshoring tasks that they are less efficient in. Productivity gains in a GVC can also be real-

ized through increased access to new, cheaper, and a greater diversity of input varieties and learn-

ing externalities  as well  as technology spillovers  (Baldwin &  Robert- Nicoud,2014). Moreover, 

participating in a GVC exposes supplier firms to intense foreign competition, creating incentives 

for them to  become more efficient and cost  competitive (Chiarvesio etal.,2010).  Further,  GVC 

firms can  benefit from  cost complementarities  from two- way trade,  resulting in  a higher  pro-

ductivity premium for two- way traders (Wagner,2012). Using a panel of  Chilean manufacturing 

firms,  Kasahara and Lapham  (2013, p. 315)  estimate that  “a firm  can save  between 7  and 26 

percent of the per- period fixed costs and  sunk costs associated with trade by simultaneously en-

gaging in both export and import activities.”

Empirical  research  on  productivity implications  of   GVCs  can  largely  be  divided  into  three 

strands. The first strand, dominant in the economics literature, uses country-  and sector- level data 

from input– output tables.  Several studies in this  strand document a positive impact  of GVCs on 

productivity (Constantinescu  etal., 2019; Kummritz, 2016; Pahl &  Timmer,2020). The  second 

strand, dominant in the  GVC  literature, uses in- depth qualitative and sectoral  case- study data to 

examine how  developing countries  can upgrade in  GVCs and  climb up  the value chain  ladder. 

There is a consensus in these studies that benefits of linking into GVCs are critically shaped by the 

type of governance structures firms are operating under (Gereffi  etal.,2005) as well as their own 

technological capabilities (Morrison et al., 2008). Firms in relational governance chains have been 

found to provide the most ideal conditions for supplier upgrading (Giuliani etal.,2005; Humphrey 

& Schmitz, 2002). These chains are  characterized by mutual dependence and tight linkages  be-

tween buyers and suppliers, which often lead to face- to- face interactions and mutual learning.

A third strand of literature exploits  quantitative firm- level data to examine productivity ben-

efits of GVCs. For Europe, Veugelers etal. (2013) find that few firms are intensively engaged  in 

GVCs, but these tend to  be larger, more trade intensive, more  innovative, and more productive. 

Baldwin and  Yan (2014) show  that 28%  of  Canadian GVC manufacturing  firms in  the period 

2002– 2006  were  generally  larger  and  more  productive and  paid  higher  wages  than  non-  GVC 

firms.  Even after  controlling  for  self- selection through  propensity  score  matching  (PSM),  the 

authors find that firms become more productive by joining GVCs. For Italian firms, the existence 

of a productivity premium is found in Giovannetti and Marvasi (2018) and Brancati etal.(2017). 

Also using PSM  for Latvian and Estonian  firms, Benkovskis etal.(2020)  find that export entry 

KEYWORDS

global value chains, productivity, propensity score matching

JEL CLASSIFICATION

F14; F15; D24

706

|

BANGA

into GVCs can increase productivity, employment, and wages. A handful of studies have further 

attempted to identify GVC governance using firm- level secondary data. Higher productivity ben-

efits from relational chains have been confirmed in the case  of Italian suppliers (Agostino et al., 

2020; Brancati et al.,2017). The literature on developing- country firms is however scant;  using 

firm- level data from the World Bank Enterprise Surveys (WBES), productivity gains from linking 

into GVCs are found for Latin America (Montalbano etal.,2018); Egypt and Morocco (Del Prete 

etal.,2017); and Indonesia, Philippines, and Vietnam (Urata & Baek,2021). Using data from the 

South African Revenue Services, Edwards  etal.(2018) also find South African  GVC firms to be 

more productive than non- GVC firms.

In this paper, we take the  case of India and  empirically examine the impact of  GVC embed-

dedness on total factor productivity (TFP) using a  large- scale unbalanced panel of  Indian man-

ufacturing f irms, constructed from  Prowess for  the period  2000/2001– 2014/2015. A  two- stage 

empirical strategy  of PSM  and system generalized  method of  moments (GMM) is  deployed to 

deal with econometric issues of self- selection and endogeneity. India presents an interesting yet 

under- researched case study; it is largely a supplier economy in GVCs, with very few lead firms 

(Ray & Miglani,2018), and even though it has manufacturing capability in most sectors, it is less 

integrated in GVCs compared to other South Asian economies (Athukorala,2013).

To our knowledge,  no study has  yet examined the causal  impact of GVC embeddedness on 

firm- level TFP for India. We make three important contributions to the existing literature. First, 

we add to the nascent  empirical literature on development implications of GVCs in India  using 

a quantitative micro- level  approach with firms as the central unit  of analysis. The closest study 

to our  research is that by Tucci (2005), which  finds a positive  impact of vertical  specialization 

(VS) on  productivity of  Indian firms in the  period 1997– 2001 using  a sample  of  Indian firms 

from the  Investment Climate Survey. Another study,  Anderson etal. (2018), notes that  Indian 

two- way traders are 15% more  productive than firms that  only export, but this analysis is based 

on difference- in- means tests.  Second, we  use TFP, constructed from  the Levinsohn– Petrin ap-

proach, as the  outcome variable, which  enables us to account  for firm heterogeneity in  capital 

intensity. Third, we quantify and analyze three aspects of  GVC embeddedness in Indian  firms: 

(1) linking into GVCs, (2) increasing embeddedness in GVCs, and (3) GVC governance. In doing 

so, we take the Indian TFP literature beyond relationships based solely on ownership structures 

to a more “network” approach, as in the GVC literature, which can help in identifying strategies 

for successfully increasing productivity gains in India.

Section 2 provides information on data sources and construction of  variables. Section 3 pres-

ents the summary and  descriptive statistics. Section 4 describes the empirical strategy  deployed 

in the paper. Section 5 presents the results, and Section 6 provides the conclusion.

2

|

DATA AND CONSTRUCTION OF VARIABLES

We construct  an unbalanced panel  of Indian manufacturing  firms, spanning the  period 2000– 

2001 (hereafter 2001) to 2014– 2015 (hereafter 2015) from the firm- level data set— Prowess, com-

piled by the Centre for Monitoring Indian Economy Pvt Ltd.1 Manufacturing industries covered 

are provided in Appendix A1. As part of data cleaning, observations with real value added to labor 

or capital  ratio below  the first  percentile or above  the 99th  percentile are removed.  Similarly, 

observations with  negative values for sales,  output, or value  added as well as firms that report 

data for  less than 3 years are excluded  from analysis. The final  panel provides information  on 

7,338 Indian firms. The number of firms reporting  data in each year varies but is approximately