Trade Liberalisation, Technology Import
and Industrial Productivity:
Evidence from Indian Manufacturing Firms
R. Rjiesh
Institute for Studies in Industrial Development
4, Institutional Area, Vasant Kunj Phase II, New Delhi ‐ 110 070
Phone: +91 11 2676 4600 / 2689 1111; Fax: +91 11 2612 2448
E‐mail: [email protected]; Website: http://isid.org.in
November 2018
ISID
Working Paper
203
© Institute for Studies in Industrial Development, 2018
ISID Working Papers are meant to disseminate the tentative results and findings obtained from the ongoing research activities at the Institute and to attract comments and suggestions which may kindly be addressed to the author(s).
CONTENTS
Abstract 1
1. Introduction 2
2. Technology Import and Productivity in Indian Manufacturing: Brief Review 4
3. Methodology and Data Sources 6
3.1 Methodology of the Study 6
3.2 Data Sources and Variable Construction 10
4. Descriptive Statistics 13
4.1 Overview of Sample Data 14
4.2 Technological Characteristics of Indian Manufacturing Firms 17
5. Estimation Results 19
5.1 Impact of Technology Import on Manufacturing Output 21
6. Concluding Remarks 24
References 26
Appendix 29
List of Figure(s)
Figure 1 Composition of Embodied and Disembodied Technology Import:
By Use‐based Classification 18
Figure 2 Technology Intensity in Indian Manufacturing: Select Indicators 19
List of Table(s)
Table 1 Distribution of Sample Manufacturing Firms: By Use‐based Classification 14
Table 2 Trend in Select firm‐level Characteristics: By Use‐based Classification 15
Table 3 Production Function Estimates for Indian Manufacturing Sector 20
Table 4 Production Function Estimates for Indian Manufacturing:
By Use‐Base Classification 21
Table 5 Productivity Determinant Model Estimation Results:
By Use‐based Classification 22
Table A1 Distribution of Sample firms Selected: By 2‐digit NIC 29
Table A2 Firms with consistent Sales and Technology Import for the Entire Period:
By Use‐Based Classification 30
Table A3 Technology Import by Top 10 firms 30
Table A4 Summary Statistics 31
Trade Liberalisation, Technology Import and
Industrial Productivity:
Evidence from Indian Manufacturing Firms
R. Rjiesh*
[Abstract: The present study is an attempt to examine the impact of technology import on productivity
of Indian manufacturing firms for the period of 1995‐2016. The study estimated an augmented Cobb‐
Douglas production function incorporating technology import, both embodied technology (capital goods
import) and disembodied technology (spending on royalties, technical fees etc.), as additional output
determinants. The production function is estimated using the semi‐parametric algorithm of Levinsohn
and Petrin (LP) that tries to correct the simultaneity bias resulting from input choices and productivity
shocks. In order to assess the robustness of the baseline model, a productivity determinant model is also
estimated using the standard panel estimation method. The empirical analysis is based on an unbalanced
sample of 4616 manufacturing firms, which are further segregated into firms that produce basic goods,
intermediate goods, capital goods and consumer goods, representing various use‐based categories of
manufacturing production. The descriptive statistics reveal the considerable expansion of manufacturing,
especially the trade linkage through import and export, during the sample period. However, some growth
deceleration is evident post‐crisis period of 2008. Most firms have found to occasionally engage in the
import of technology and the preferred mode of acquisition is embodied technology. Compared to the
internal technological effort (domestic R&D), the preference is towards acquiring technology from
abroad. The consumer goods and intermediate producing segments are the major consumers of foreign
technology inputs. The econometric results based on the direct estimation of the production function and
the productivity determinant model for the entire manufacturing, as well as use‐based categories largely
substantiates the conjecture that technology import has a positive and statistically significant impact on
Indian manufacturing output. This supports the predictions of new trade and growth theories that
international trade is a significant vehicle of technology transfer that provides dynamic productivity
benefits in manufacturing.]
Keywords: Technology Import, Productivity, LP, Indian Manufacturing, Panel Regression
JEL Classification: F14, D24, L6, O33
* Assistant Professor, ISID, New Delhi; E‐mail: [email protected]
Acknowledgement: The study is a part of the project sponsored by ICSSR. I duly acknowledge the
financial support and assistance by the ICSSR. I thank the institutional support of ISID during the
various stages of my research work. Some part of the work is also based on my Ph.D. for which I
am indebted to Prof. Alokesh Barua and Prof. K.L. Krishna. I thank the valuable comments and
suggestions that I received during the internal seminar held at ISID. However, any remaining
errors are solely my responsibility.
2
1. Introduction
Since the celebrated empirical work of Kaldor (1967), the critical role of industrialisation in
the process of economic growth is well recognised. It is argued that the allocation of
economic resources towards industrial activities often induces structural change bonuses
in terms of higher productivity, learning opportunities, technological progress and
spillover effect to the rest of the economy (Szirmai, 2013). Industrial growth arguably
involves considerable investment in production technology and innovation that not only
improve cost efficiency and quality, but also enable the less‐income economy to catch‐up
with the frontier economies of the world. The neo‐classical growth theories, especially
since the seminal contribution by Solow (1957), have clearly pointed out the significance
of the technical change in augmenting the growth process of the economy. The recent
growth theories, while relaxing the simplified notion of technology as an exogenous shock
or black box, regard it as an outcome of deliberate economic activity governed by economic
incentives. In this framework, an economy’s growth rate is determined not only by the rate
of capital accumulation, but also by the rate of investment in knowledge and skill intensive
innovations (see of Grossman and Helpman (1991); Rivera‐Batiz and Romer (1991) among
others).
However, it is well known that countries are not equally endowed with the stock of
knowledge, as the bulk of innovation activities are concentrated in the developed countries
(Rijesh, 2015). In general, the low‐income countries are characterised by structural
rigidities including market imperfections, distortions and lack of incentives to invest in
risk‐bearing innovative activities. Rather than investing in new technology generation, it
is more economical for them to access the pool of knowledge available in the rest of the
world and adapt them to the local manufacturing conditions. In this context, the new trade
and the endogenous growth theories argue that international trade can act as a vehicle of
knowledge transfer from rich countries to the poor countries (Zanello et al., 2016). In
particular, the trade‐induced dynamic effects on growth are channelled through the import
of technology‐intensive capital goods inputs which enhance the productivity and growth
of firms and industry over time. In this context, participation in international trade is
expected to benefit the low‐income countries to develop and modernise their
manufacturing sector.
The theoretical framework of Quah and Rauch (1990), Grossman and Helpman (1991) and
Rivera‐Batiz and Romer (1991) have shown that trade not only increases, but also sustains
the rate of economic growth through wider access to essential and efficient foreign inputs
like capital and intermediate goods. Since foreign capital goods are embodied with R&D
of the advanced countries, the efficiency of capital accumulation and productivity in
manufacturing improves, leading to further dynamic benefit for the entire economy (Lee,
1995; Herrerias and Orts, 2011)1. As mentioned earlier, the import of technology inputs
1 In a simple open economy model in the framework of endogenous growth theories, Rebelo (1991)
3
helps firms to reduce their marginal cost of production compared to autarky. As
production cost falls, firms will be able to devote a larger share of expenditure on
upgrading technology, and new product or process innovations thus increasing the
product scope (Damijan et al., 2014). Availability of new vintage and quality capital inputs
provide further dynamic learning effects (Bas, 2012). Generally foreign machines consist
of a bundle of ‘knowledge’ in the form of blueprints, installation support, quality control
software, and services of trained engineers and supervisors that help increase productive
efficiency in the short run and stronger domestic absorptive capacity in the long run (Mody
and Yilmaz, 2002)2.
Given the central role of technology import in industrial growth and development, it is
important to study its likely impact on improving manufacturing productivity, especially
in the less advanced countries of the world. In this context, the present study attempts to
examine the impact of foreign technology purchases on the productivity of the Indian
manufacturing firms in the recent period. The Indian case is particularly relevant, as the
economy has undergone some major changes in respect of policy in the recent decade. The
economic liberalisation policy reforms in the early 1990s have shifted the focus from highly
controlled, state‐led inward‐oriented development regime to a largely market‐oriented
outward‐oriented trade regime. Since the initiation of the comprehensive liberalisation
package, trade barriers have fallen drastically and economic integration has increased
considerably3. Due to the relaxation and lesser state‐level control over production matters,
it has become relatively easier for firms to access superior technology inputs from abroad.
For example, the share of capital goods imports in total merchandise imports—a proxy for
embodied technology—hovered around 16–23 per cent, which reflects the increased
demand for technology‐intensive inputs in the manufacturing sector. Since foreign
sourced factor inputs are embodied with better technology, knowledge bundle and
quality, incorporating or adapting them to the production process is expected to create
dynamic productivity benefits over time. In this context, the present study examines the
impact of technology import on the productivity of Indian manufacturing, using the
disaggregate data at the firm level. As a novel contribution, the present study not only
evaluates the impact of technology import on manufacturing, but also assess the impact
across different categories of manufacturing sector. In this regard, we categorised
manufacturing into four use‐based categories, namely basic goods, intermediate goods,
capital goods and consumer goods (both durable and non‐durable) producing sectors.
has shown that the growth rate is higher in a country that uses imported inputs relatively more
than domestically produced inputs for investment. 2 However, since the adoption of technology to domestic market is costly and market distortions
are prevalent in less income countries, the efficient diffusion of foreign inputs depends upon
factors such as the technological effort (Lall, 1992) and the absorptive capacity of the recipient firm
(Cohen and Levinthal, 1990). 3 For instance, the level of openness, measured as merchandise trade share of GDP has shot up from
13 per cent in 1991 to 42 per cent in 2013.
4
The rest of the paper is organised as follows. Section 2 provides a brief account of the
existing empirical studies on technology import and productivity in Indian
manufacturing. Section 3 describes the econometric methodology and data sources used
in the study. The descriptive statistical analysis of manufacturing firms’ performance and
technology characteristics are given in Section 4. Section 5 discusses the empirical results
and findings. Section 6 provides concluding remarks.
2. Technology Import and Productivity in Indian Manufacturing:
Brief Review
Since the beginning of economic planning, the importance of building a diverse industrial
base with advanced technology has been well recognised by the policymakers. However,
since the private sector’s R&D expenditure is extremely low in India, the preferred mode
of technology acquisition is from the external sources (Lall, 2001). One of the earlier studies
in this context is by Subrahmanian and Pillai (1976). The authors find that for the majority
of firms in the engineering, pharmaceuticals and dyestuff sectors, the varying degree of
foreign collaborations did not culminate in a better export performance. Kumar (1985), by
reviewing the literature on technology imports, highlights the excessive cost of such
arrangements in India. Katrak (1997) found that for 82 firms from the electrical and
electronics industries, the initial technological capability are weakly correlated to
technology import. Kumar and Siddharthan (1994) did not find any significant impact of
technology import on export performance across firms. However, most of these earlier
studies have overlooked the impact of technology import on productivity. Furthermore,
these studies cover a period of highly restrictive trade and industrial regime in which
severe restrictions on the external mode of technology purchases were the norm.
Basant and Fikkert (1996), using panel data on 787 Indian firms from 1974–75 to 1981–82,
studied the effect of foreign technology purchase, and domestic and international spillover
on Indian firms. The results indicate that when all of the industries are examined together,
there are high returns to expenditure on technology purchase and R&D. The study
suggested that India’s restriction on technology licensing agreements imposed a
substantial cost on the economy. Overall, the study finds that technology purchase by
Indian firms through licensing has a positive effect on productivity. Kathuria (2001) tested
whether the presence of foreign firms and disembodied technology imports results in
higher productivity growth for domestic firms. Using the stochastic frontier method for
368 medium‐ and large‐sized Indian manufacturing firms for the period 1975–1989, the
author found that for all manufacturing firms there was a positive effect of disembodied
technology import on productivity growth. In the case of non‐FDI firms, it was found that
embodied technology import improved productivity, whereas disembodied technology
import had a positive but not a statistically significant effect.
Hasan (2002) estimated production function using a panel data of 286 public limited
manufacturing firms from 1976–87 to determine the extent to which productivity has been
5
influenced by imported as well as domestic technologies. The imported technology was
measured through the import of capital goods (embodied technology) and supply of
design, spending on blueprints, technical assistance, royalty, etc., (disembodied
technology). The Instrumental Variable (IV) estimation of the production function reveals
a significant impact of embodied and disembodied technology purchases on firm
productivity. The disaggrete level analysis revealed that the import of disembodied
technology had a significant positive impact on productivity across technology‐intensive
sectors, like electrical & non‐electrical machinery, but not in non‐scientific firms. The
import of capital goods improved the productivity of scientific firms such as chemicals,
pharmaceuticals, and electrical and non‐electrical machinery.
Parameswaran (2009) examined the effect of international trade on R&D spillovers and
productivity of the manufacturing sector for the period 1989–2001. Using an unbalanced
panel of 2100 firms from PROWESS database, the study examined the impact of embodied
and disembodied purchase of technology on firm’s productivity. The contribution of
import of technology was assessed by estimating an augmented Cobb‐Douglas production
function using Levinsohn and Petrin (2003) methodology. The results suggested that the
effect of technology import (both embodied as well as disembodied technology) was
significant for manufacturing firms as a whole. The study finds that capital goods imports
were significant for technology‐intensive sectors, whereas the disembodied technology
was found to be significant for low technology industries. The study highlights that
technology import substitutes domestic R&D efforts. Rijesh (2015) examined the impact of
technology import on productivity performance of a large number of manufacturing firms
for the period 1995 to 2010. The LP estimates of production function for 5551 firms revealed
that technology import is a significant productivity determinant for overall manufacturing.
The technology‐intensive classification of firms revealed that the impact of capital goods
is much more pronounced across the medium technology intensive segments such as
machinery, chemicals, transport equipment, etc.
Goldberg et al. (2010a) found that trade reforms have substantially increased the access to
new types of intermediate inputs from abroad for 2927 firms during 1989–97. During the
period of liberalisation, two‐thirds of the increase in imported inputs occurred in products
that were not previously imported. Moreover, the authors observed substantial static gains
from trade through access to new imported inputs as these new varieties generated an
additional annual 4.7 per cent decline in the imported input price index which again
enhanced firms’ ability to manufacture new products. Similarly, Topalova and
Khandelwal (2011) revealed that a reduction in tariff protection improved firm‐level
productivity. Efficiency improvement results from increased competitive pressure from
lower output tariffs and access to superior foreign inputs made possible by lower input
tariffs. The study estimated industry level production function using LP methodology for
a sample of the firm during the period 1987–2001.
The brief review reveals that technology import has been a major instrument in enhancing
the productivity performance of manufacturing in India. Even though most results are
6
from a period when substantial restrictions were placed on firms’ choices on technology
acquisition, the findings of a statistically positive and significant impact is noteworthy.
Post‐reform studies have also pointed out the importance of imported technology, but
there is not much agreement on the relative merit of these purchases. That is, there are not
many evidences on the relative merit of embodied and disembodied technology import on
productivity. To the best of our knowledge, previous studies have not explored the impact
of technology import on different use‐based classifications of manufacturing. It is in this
context that the present study examines the impact of technology purchases from abroad
on firm‐level productivity of Indian manufacturing for the period 1995–2016. The
classification of firms across different use‐based product groups is expected to provide
insights into the relative impact of different modes of technology transfer in
manufacturing.
3. Methodology and Data Sources
3.1 Methodology of the Study
In this study, the impact of technology import from abroad on domestic manufacturing
output is examined through the estimation of production function at the firm level. As an
alternative framework, the study has also estimated a productivity determinant model,
popularly known as the two‐step procedure, incorporating technology import and domestic
technological capability.
(a) Production Function Estimation
The impact of technology import on manufacturing output is examined with the help of
firm level production function estimation using the semi‐parametric estimation algorithm
proposed by Levinsohn and Petrin (2003) or LP in short. The LP has emerged as the
standard procedure to derive productivity at the firm level in empirical literature,
especially for the Indian manufacturing sector4. The LP method is a semi‐parametric
estimation procedure that tackles the issue of endogeneity involved with input choices of
firms in the production procedure. Since the seminal contribution of Marschak and
Andrews (1944), it is well known that identification of productivity parameters from the
estimation of production function is problematic due to the potential correlation between
factor inputs and productivity. Most often, the firms input choices are determined by
characteristics of firm or efficiency. This brings the issue of simultaneity due to the
potential correlation between the choice of input mix and productivity shocks that are
unobservable to the econometricians but are known to the firm (De Loecker, 2007). This
4 Several recent empirical studies has used the Levinsohn and Petrin (2003) or LP methodology in
short, for estimating production function at the firm level. Some of the prominent studies being
Parameswaran (2009), Rijesh (2015), Thomas and Narayanan (2016) among others.
7
implies that firms’ choices of inputs are subject to the perception of its own productivity
performance (Rijesh, 2015). In such cases, the OLS estimates of the production function
would be biased because of the potential correlation between productivity and level of
inputs5. An innovative way to correct for the simultaneity bias is the semi‐parametric
method developed by Olley and Pakes (1996) and later modified by Levinsohn and Petrin
(2003).6
The LP method incorporates the unobservable effects on productivity from inputs and
produces reliable input coefficients. In order to address the endogeneity problem, the LP
methodology uses intermediate inputs such as raw material or energy as a proxy variable
which is assumed to have a monotonic relationship with the unobservable firm‐specific
productivity shock. The assumption of monotonicity allows inverting the intermediate
input demand function to unearth the unobservable productivity term as a non‐parametric
function of the intermediate input and capital. Thus, the insight of the LP method is that
the observable characteristics of the firm can be modelled as a monotonic function of the
productivity of the firm. Inverting such a function allows modelling the unobserved
component of firm productivity as a function of the observed variables. Therefore, in the
present study we follow the L‐P estimation procedure for estimating the impact of
technology import on manufacturing output. We use an augmented Cobb‐Douglas
production function that includes technology import purchases as additional factor
inputs.7 The details are as follows.
Under the production function approach, it is assumed that firm’s output (Y) depends on
the use of traditional inputs such as capital (K) and labour (L) subject to the state of
technology (T). We consider T to be augmented by the adoption of foreign technology (FT)
purchases in the form of capital good equipment (embodied technology import, ETI) and
payment for licensing involving the purchase of design, blueprints and technical assistance
by paying royalties and technical fees (disembodied technology import, DETI). Assuming
a Cobb‐Douglas production function, we can write
𝑌 𝐴 𝐾 𝐿 𝑀 𝐸𝑇𝐼 𝐷𝐸𝑇𝐼 ‐‐‐ (1)
where, 𝑌 refers to output of the firm i in period t, 𝐴 is the efficiency level of firm i in
period t, 𝐾 , 𝐿 , 𝑀 , 𝐸𝑇𝐼 , 𝐷𝐸𝑇𝐼 are inputs of capital, labour and material,
embodied technology and disembodied technology, respectively. The natural log
of equation (1) gives a production function (equation 2) that is linear in logs of
output and inputs.
5 Another potential problem from OLS estimation of production function is the issue of selection
bias. This result from the decision of the firm to continue to serve the market based on its perceived
productivity or expected future profitability. 6 For a detailed outline on Semi‐parametric methods, see Yatchew (2003). 7 This is similar to the methodology adopted by Parameswaran (2009) and Rijesh (2015).
8
𝑦 𝑏 𝑏 𝑘 𝑏 𝑙 𝑏 𝑚 𝑏 𝑒𝑇𝐼 𝑏 𝑑𝑒𝑇𝐼 𝜀 ‐‐‐ (2)
where, 𝑦 denotes log of gross output of the firm, k is the log of capital stock, l is the log
of labour input, m is the log of raw material consumed, eT is the log of embodied
technology, deT is the log of disembodied technology and is the error term. Here,
ln 𝐴 𝑏 𝜀 ; where b measures mean efficiency level across firms and over
time and ε is the time‐producer‐specific deviation from that mean (Van Beveren,
2012). 𝜀 can be decomposed into 𝜀 𝜔 𝛾 where 𝜔 is the firm‐specific
productivity difference not captured by explanatory variables and 𝛾 is a pure
random variable capturing unexpected deviation from mean due to measurement
error, unexpected delays or other external disturbances (Van Beveren, 2012 and
Rijesh 2015). The firm‐specific shock (i.e., 𝜔 ) is a state variable that is observable
only to the firm manager and influences the input demand choices. Incorporating
them, we can re‐write equation (2) as
𝑦 𝑏 𝑏 𝑘 𝑏 𝑙 𝑏 𝑚 𝑏 𝑒𝑇𝐼 𝑏 𝑑𝑒𝑇𝐼 𝜔 𝛾 ‐‐‐ (3)
In order to control the variation in input that are correlated with the unobserved
productivity shock, the LP method uses raw material expenses of the firm as the control
variable. The assumption here is that firms’ material demand function is monotonically
increasing in its unobserved productivity, conditional on 𝑘 , a state variable. That is 𝑚𝑚 𝜔 𝑘 . Assuming monotonicity, the inverted raw material demand function can be
written as 𝜔 𝜔 𝑚 𝑘 , which shows that the unobservable productivity component
depends upon observable inputs and hence control for 𝜔 in the production function
estimation. Incorporating the productivity shock proxy, we can rewrite equation (3) as
𝑦 𝑏 𝑙 𝑏 𝑒𝑇𝐼 𝑏 𝑑𝑒𝑇𝐼 𝜃 𝑚 𝑘 𝛾 ‐‐‐ (4)
where, 𝜃 𝑚 𝑘 𝑏 𝑏 𝑘 𝑏 𝑚 𝜔 𝑚 𝑘 .
The estimation of production function takes place at two stages and the details are
elaborated in Levinsohn and Petrin (2003). The production function estimation, which is a
direct evidence of the impact of technology import on productivity, is carried out for 4616
manufacturing firms over 1995–2016 period. The disaggregate performance of firms is
analysed by estimating the production function across five use‐based classifications of
firms, namely basic goods, intermediate goods, capital goods, consumer durable goods,
and consumer non‐durable goods‐producing segments.
(b) Productivity Determinant Model
In the methodology outlined in the sub‐section 3.1(a), we propose to examine the impact
of technology import on output by estimating the augmented Cobb‐Douglas production
function. This is the direct method of assessing the impact of inputs on productivity. An
9
alternative method is to derive productivity residual from the standard production
function and model the determinants including technology import. Moreover, technology
accumulated from abroad also requires firms to have considerable internal technological
capability to assimilate, master and adapt the technology internally, which is often
reflected in their expenditure on R&D. In order to examine the impact of both external
technology purchases and domestic technological capability on the productivity of firms,
we model the determinant of productivity using the standard panel estimation
techniques8.
In this case, the first step is to estimate the firm‐level production function. We estimated
the Cobb‐Douglas production function (equation 5) using LP (2003) methodology.
𝑞 𝑏 𝑏 𝑘 𝑏 𝑙 𝑏 𝑚 𝜔 𝛾 ‐‐‐ (5)
This allows us to calculate productivity or Total Factor Productivity (TFP) of the firms by
taking the difference between actual and predicted values of manufacturing output (see
equation 6).
𝑇𝐹𝑃 𝑞 𝛽 𝑘 𝛽 𝑙 𝛽 𝑚 ‐‐‐ (6)
In the second step, productivity is modelled to be a function of technology and other
relevant factors. That is, we regress the firm’s productivity on the import of technology (TI,
which consists of ETI and DETI), Research and Development (RD) and exports (EXP). In
this case, the RD represents the domestic technological effort and capability of the firm
while the export variable represents the nature of outward orientation. Specifically, we
estimate the following equation,
𝑙𝑛𝑇𝐹𝑃 𝛼 𝛽 𝑙𝑛𝐸𝑇𝐼 𝛽 𝑙𝑛𝐷𝐸𝑇𝐼 𝛽 𝑙𝑛𝑅𝐷 𝛽 𝑙𝑛𝐸𝑋𝑃 𝜁 𝜏 𝜖 ‐‐‐ (7)
where, 𝑙𝑛𝑇𝐹𝑃 is the productivity of firm i in period t, ETI is embodied technology
import, DETI is disembodied technology import, RD is real expenditure on
Research & Development, EXP is real exports. 𝜁 and 𝜏 are coefficients of industry and year dummies. All variables are in natural logs. The data sources and variable
construction details are given in the sub‐section 3.2.
8 The two‐step procedure is the most popular method to assess the impact of technology or other
determinant on productivity. However, it has to be noted that this procedure often produces
inefficient and potentially biased coefficients because of the potential correlation between the
productivity determinant factors and production factor variables used to derive the residual in
the first stage of production function (see Harris, 2005). Therefore, the direct estimation of the
production function is relatively superior in exploring the impact of technology on productivity.
This is the main drawback of the two‐stage method explored in this section.
10
3.2 Data Sources and Variable Construction
The main source of data used for the empirical analysis is Prowess IQ version 1.81. Prowess
is an online client software with an interactive querying system maintained by the Centre
for Monitoring Indian Economy (CMIE) Pvt. Ltd. CMIE collects firm wise data from the
balance sheets and income statements of the company and covers the financial
performance of all listed (stock market) and a large number of unlisted companies of India.
The financial performance indicators are based on the audited Annual Financial Reports
of companies and information submitted to the Ministry of Company Affairs, company
filings with stock exchanges and prices of securities listed on the major stock exchanges.
Apart from the large cross‐section of manufacturing industries, the database covers
services, utilities and financial industries as well. The Prowess IQ covers 60–70 per cent of
the organised manufacturing sector, 75 per cent of corporate taxes, and 95 per cent of excise
duties collected by the Government of India (Goldberg et al., 2010b). Apart from
fundamental financial performance indicators, the Prowess database provides detailed
information on companies spending on technology from abroad such as import of capital
goods and foreign exchange spending on royalty, technical fees, etc., which is the main
rationale for using the database for our study.
From the Prowess database, we extracted selected Annual Financial Statement indicators
of companies such as sales, compensation to employees, gross fixed capital formation, raw
material expenditure, exports, imports, R&D expenditure, etc., belonging to the organised
manufacturing sector. We followed the National Industrial Classification (NIC9) scheme of
2008, according to which manufacturing belongs to NIC10 (food products) to NIC32 (other
manufacturing). All reported data were identified at the 5‐digit level of NIC for the period
1995 to 2016. The compiled data is cleaned before the final selection and the procedure
involves several steps. To begin with, we dropped 8592 firms that do not import
technology in the form of purchase of capital goods or payment of royalty, technical fees,
etc., for any of the reference period. For the remaining firms, we removed firms that do not
report sales (46 firms), gross fixed capital (3 firms), wages (6 firms), and material (45 firms).
For the remaining, we decide to remove firms with less than five years of sales data (1080
firms) during the study period. Finally, the sample data consisted of 4616 manufacturing
firms. The detailed list of manufacturing firms by 2‐digit NIC 2008 is provided in Table A1
in the Appendix. We can observe that the largest number of firms belong to the chemical
sector (556) while the smallest is in the printing and reproduction of recorded media (6).
Alternatively, the selected sample firms were further categorised into use‐based
classification which is given in Table 1 in section 4. We can observe that the consumer good
producing firms have the larger presence in the sample whereas the basic good firms have
the lowest presence. Based on the sample data, the real variables required for the empirical
analysis is constructed and the details are as follows.
9 NIC is based on ISIC nomenclature of the UN. The NIC‐2008 follows the broad structure of ISIC
rev4.
11
Output (Y): We have taken the sales of goods by firms to represent the output variable.
The reported nominal sales figures were deflated by the Wholesale Price Index (WPI) of
manufacturing product group at the disaggregate level to derive the value of output in
2004–05 prices. In case where the appropriate deflators are not available, the price index
corresponding to the nearest product group is utilized. The WPI index is available from
the office of Economic Advisor, Government of India website. For the entire period of
study, there were two WPI series i.e., WPI at 1993–94 prices and WPI at 2004–05 base prices.
We selected the common product groups among them and spliced the series to derive the
WPI at the base year 2004–05=100. As far as possible, we selected the relevant WPI at the
lowest levels of disaggregation to avoid any plausible bias in the estimation.10
Labour Input (L): Since the database does not report direct employment figures, we define
labour input in value terms, i.e. compensation to employees at 2004–05 prices. This figure
includes all cash payments in kind made by a company to its employees. We use the
Consumer Price Index (CPI) for industrial workers to deflate the nominal values of wages
at 2004–05 prices. This practice has been common among several past studies on
productivity in Indian manufacturing. For instance, see Srivastava (1996), Kambhampati
(2003), Mazumdar et al. (2009), Kim and Saravanakumar (2012), Rijesh (2015) among
others11.
Capital Input (K): In this study, Gross Fixed Capital Assets (GFA) expressed in 2004–05
measures the capital input. The GFA includes both tangible assets such as land, building,
plant, and machinery, and intangible assets such as goodwill assets, software, etc. Since
the reported GFA is the book value of the asset at current prices, we use the Perpetual
Inventory Method (PIM) to estimate the stock of capital at the constant price for each firm
in the sample. The PIM revalues the book value of capital, which is given at historical cost
to a replacement value for a given base year. The procedure is based on the methodology
proposed by Srivastava (1996) and later adopted by many studies including Balakrishnan
et al. (2000), Parameswaran (2009), Rijesh (2015), Thomas and Narayanan (2016), among
others. This method revalues the capital given at historical cost to a base year. The
calculation of capital stock is briefly explained below.12
The PIM method requires that the base year capital stock (𝐾 has to be revalued at the
replacement cost. The reported GFA in the database is converted to the replacement of cost
of the base year capital (i.e. GFA in the year 1995) using the Revaluation factor (𝑅 ). The
𝑅 is estimated through the procedure outlined in Srivastava (1996). It is assumed that the
life of plant and machinery is 20 years. For firms incorporated before 1975–76, it is assumed
that the earliest vintage of the machine in their capital mix dates back to the year of
10 In the absence of price data at the firm level, researches usually use the price series at the industry
level. If the firm price series are correlated with input choices of firms, the use of industry‐level
price will bias the input coefficients. For further discussion, see Van Beveren (2012). 11 A plausible issue with this measure is the potential bias in estimation where workers shares the
benefits of productivity (Sivadasan, 2006). 12 For detailed explanation, see Srivastava (1996).
12
incorporation. For all firms, we assume that the price of capital () has changed at a constant rate from 1975 or the year of incorporation, whichever is later, up to 1995. That is:
𝜋 1 ‐‐‐ (8)
In addition, we assume that investment has increased at a uniform rate for all firms.
𝑔 1 ‐‐‐ (9)
To derive 𝜋, we use the capital formation price series derived from the Gross Fixed Capital
Formation series, current and constant price for the organised manufacturing collected
from NAS, published by the CSO. The rate of growth of investment 𝑔 is approximated
by the rate of growth of gross fixed capital formation in manufacturing at 2004–05 prices.13
Assuming that the capital stock of the earliest vintage is t period old, the 𝑅 is derived
using the formula (10)
𝑅 ‐‐‐ (10)
The replacement cost of GFA at firm level is derived by multiplying the balance sheet value
(historical cost) of GFA in 1995 by 𝑅 for each firm in the sample. The capital stocks for the
subsequent time period were calculated by adding the gross investment (GI) to the earlier
stock. GI is obtained from the relation
𝐺𝐼 𝐺𝐹𝐴 𝐺𝐹𝐴 ‐‐‐ (11)
The real capital stocks (K) is obtained by deflating the revalued GFA by the capital good
price series (at 2004–05 prices) constructed using gross capital formation series (current
and constant) obtained from the National Accounts Statistics.
Material Input (M): The raw material expenses valued at 2004–05 prices represent the
material input variable. This includes the value of raw materials, power and fuel
consumptions. Raw materials are the sum of expenses on raw materials, stores, spares and
tools while power and fuel consumption expenditure includes expenses on power, fuel
and water. The nominal value of the raw material cost was deflated using raw material
price indices, base 2004–05 = 100. Since the price indices for raw materials are not readily
available, we constructed it using weights obtained from Input‐Output Transaction Matrix
Table, published by the CSO and the product price indices from the Index of Wholesale
Prices in India, base 2004–05 = 100. The price series is a weighted index of the WPI of major
13 See Rijesh, (2015) for similar exposition.
13
input groups, where the weights for the respective industries are calculated from the
absorption matrix of input‐output transactions.14
Technology Import Input (TI): For a less developed country such as India, a major means
of acquiring technology is through imports. The import of technology mainly consists of
two activities. One is through the purchase of machinery and equipment embodying better
technology and the second is the royalties, licensing, and technical fees paid by the firms
for foreign technology. The former reflects technology that is embodied in capital goods and
the latter indicates the disembodied form of technological flows from abroad. Using the
Prowess IQ database, we collected each company’s import of capital goods and payment
towards royalties, technical fees, etc. The nominal values of import of embodied
technology were deflated by the Unit Value Index (UVI) of machinery & equipment
import, base 2004–05=100. The disembodied technology purchases in real term are
constructed using the R&D deflator of the US, with the assumption that the US has been a
major provider of technical assistance to the Indian firms.15 The deflator is constructed
using the current and constant series of R&D for the manufacturing sector of US and
adjusted for the Rupee‐Dollar exchange rates. The total import of technology, both
embodied and disembodied technology, were deflated using the average price series of
R&D deflator and capital good import price, base 2004–05=100.
Research and Development (RD)
The expenditure on Research and Development (R&D), both capital and current account,
is taken to construct the real R&D expenditure (RD) expressed in 2004–05 prices. The
nominal values were deflated using R&D deflator which is constructed by taking the
average of capital good prices and consumer price index of industrial workers. The data is
collected from NAS and the Ministry of Labour & Employment, Government of India. The
RD measures the internal technological capability of the firm in the sample.
Real Exports (EXP)
The database provides information regarding firms selling goods abroad. The nominal
value of exports is deflated with the Unit Value Index (UVI) of exports at the 2‐digit level.
This provides real export (EXP) of manufacturing firms at 2004–05 prices.
4. Descriptive Statistics
In this section, we provide the basic characteristics of the sample data using standard
statistical tools and methods.
14 The construction of the price index follows the method outlined in Rijesh (2016). 15 This is similar to Basant and Fikkert (1996), Hasan (2002), Parameswaran (2009), and Rijesh (2015).
14
4.1 Overview of Sample Data
The firm wise data that we collected from the PROWESS IQ database were identified
across 5‐digit National Industrial Classification (NIC) 2008 registry, prepared by CSO,
Government of India. The NIC is the standard classification for industry product groups
in India which adopts framework of ISIC of the UN. The disaggregate industry
classification of firms products allowed us to categorise firms according to the specific
use for which the products are developed in the market. We adopted the Index of
Industrial Production (IIP) 2011–12 classification that clubs NIC‐2008 product groups
into six use‐based categories, namely Primary goods, Infrastructure/construction goods,
Intermediate goods, Capital goods, Consumer durables, and Consumer non‐durables.
Using the earlier IIP series of 2004–05, we standardised the use‐based product groups
into five main heads, namely Basic Goods (BAG), Intermediate Goods (IMG), Capital
Goods (CAG), Consumer Durable Goods (CDG), and Consumer Non‐Durable Goods
(CNDG).
The BG includes raw industrial products such as starch, coal, kerosene, basic chemicals,
fertilizers, metals, iron & steel, and aluminium. The IMG covers a wide range like textile
yarns, jute, tanned leathers, wood materials, refined petroleum products, plastic products
in primary forms, synthetic rubber, chemicals, insulated wires, accumulators & cells, and
railway instruments. The CAG is a durable producer good used to produce other goods
and includes a variety of machinery equipment and tools such as lathes, drills, agriculture
machinery, textile machinery, cement machinery, electric motors, and transport
equipment. The consumer goods industry, which consists of both CDG and CNDG,
includes final consumption (end) based products of food, beverages, leather, rubber,
electrical and non‐electrical appliances, automobiles, and transport vehicles. The
distribution of firms according to the use‐based classification is given in Table 1.
Table 1
Distribution of Sample Manufacturing Firms: By Use‐based Classification
Use‐based classification Number of firms (%)
Basic goods 144 (3.1)
Intermediate goods 1746 (37.8)
Capital goods 793 (17.2)
Consumer durables 869 (18.8)
Consumer non‐durables 1064 (23.1)
Manufacturing Sector 4616 (100.0)
Source: Author classification of Prowess data (CMIE).
We can notice that a large majority of firms engage in producing consumer goods that
include both durable and non‐durable products (42%) followed by intermediate goods
(38%), capital goods (17%) and basic goods (3%). In terms of broad sectors, the largest
amounts of firms (556) were under chemicals and chemical product groups (NIC code 20).
However, we do find that not all firms consistently engaged in production activities and
15
purchase of technology during the entire period of study. For instance, only 1005 firms
were reported to have consistent sales throughout the study period. Moreover, only 194
firms were engaged in technology purchases for the entire period from 1995 to 2016 (See
Table A2 in the appendix for the detailed list based on use‐based distribution of firms in
the sample).
In Table 2, we provide the growth trend among select firm‐level characteristics for the
entire period of study (1995–2016) as well as for two sub‐periods (1995–2007 and 2008–
2016). The first and second periods correspond to the pre‐ and post‐global financial crisis,
respectively. As before, the entire manufacturing is classified across five use‐based
classifications. The selected firm‐level characteristics are sales, wages, capital, exports,
imports, embodied technology import, disembodied technology import, technology
import, and domestic R&D expenditure. All variables are deflated by appropriate price
indices having 2004–05 as base year.
Table 2
Trend in Select firm‐level Characteristics: By Use‐based Classification (%)
BEC Variables 1995–2007 2008–2016 1995–2016
Basic goods
Sales 8.1 (6.4) 7.4 (5.6) 7.8 (6.1)
Wages 3.3 (6.2) 6.0 (6.0) 4.4 (6.1)
Capital 9.2 (9.4) 10.2 (8.8) 9.6 (9.2)
Exports 12.9 (2.1) 4.3 (1.3) 9.2 (1.8)
Imports 7.4 (6.5) 18.9 (4.1) 12.3 (5.5)
Embodied Technology import 51.5 (3.9) ‐8.6 (3.9) 25.7 (3.9)
Disembodied Technology import 15.1 (7.1) 25.1 (3.0) 19.4 (5.4)
Import of Technology 31.5 (4.5) ‐2.1 (3.8) 17.1 (4.2)
R&D Expenditure 5.5 (3.3) 10.7 (1.1) 7.7 (2.4)
Intermediate goods
Sales 8.7 (56.7) 4.1 (55.3) 6.7 (56.2)
Wages 4.8 (42.8) 2.6 (37.1) 3.8 (40.5)
Capital 10.9 (61.5) 11.2 (58.5) 11.0 (60.3)
Exports 16.8 (53.7) 4.8 (62.4) 11.7 (57.2)
Imports 12.4 (7.6) 0.9 (6.9) 7.5 (7.3)
Embodied Technology import 6.5 (56.8) 4.2 (55.3) 5.5 (56.2)
Disembodied Technology import 8.3 (45.2) ‐2.5 (13.0) 3.7 (32.0)
Import of Technology 8.9 (55.0) 4.8 (47.8) 7.1 (52.1)
R&D Expenditure 14.5 (20.7) 9.6 (12.1) 12.4 (17.2)
Capital goods Sales 12.4 (11.5) 6.8 (12.4) 10.0 (11.9)
Wages 4.7 (22.5) 3.9 (22.4) 4.3 (22.5)
Capital 10.6 (8.8) 13.8 (10.3) 11.9 (9.4)
Exports 10.2 (9.5) 1.4 (7.2) 6.4 (8.6)
Imports 15.1 (9.5) 4.5 (8.1) 10.5 (8.9)
16
BEC Variables 1995–2007 2008–2016 1995–2016
Embodied Technology import 13.1 (10.8) ‐7.1 (12.6) 4.5 (11.5)
Disembodied Technology import 12.7 (13.7) 14.3 (25.8) 13.4 (18.7)
Import of Technology 15.4 (11.2) ‐2.6 (14.8) 7.7 (12.6)
R&D Expenditure 19.0 (27.9) 10.0 (33.9) 15.2 (30.3)
Consumer durables
Sales 14.7 (11.1) 10.9 (12.5) 13.1 (11.7)
Wages 7.4 (11.8) 6.8 (14.0) 7.2 (12.7)
Capital 12.0 (9.4) 11.6 (9.9) 11.9 (9.6)
Exports 12.7 (12.7) 6.0 (11.8) 9.8 (12.3)
Imports 12.4 (6.7) 7.2 (6.2) 10.2 (6.5)
Embodied Technology import 16.7 (19.7) ‐2.4 (17.7) 8.5 (18.9)
Disembodied Technology import 28.7 (26.2) 22.6 (46.9) 26.1 (34.7)
Import of Technology 19.6 (20.7) 3.8 (23.0) 12.8 (21.6)
R&D Expenditure 12.0 (16.7) 14.5 (12.9) 13.1 (15.2)
Consumer non‐durables
Sales 9.3 (14.2) 9.1 (14.1) 9.2 (14.2)
Wages 7.8 (16.6) 6.8 (20.4) 7.4 (18.2)
Capital 12.9 (10.8) 12.5 (12.5) 12.7 (11.5)
Exports 16.4 (22.0) 6.0 (17.4) 11.9 (20.1)
Imports 14.8 (69.8) 4.7 (74.7) 10.4 (71.8)
Embodied Technology import 8.8 (8.8) 2.0 (10.5) 5.9 (9.5)
Disembodied Technology import 13.6 (7.8) 23.8 (11.3) 18.0 (9.2)
Import of Technology 10.7 (8.6) 4.7 (10.7) 8.1 (9.5)
R&D Expenditure 21.0 (31.3) 19.6 (39.9) 20.4 (34.9)
Manufacturing
Sales 9.6 (100.0) 6.1 (100.0) 8.1 (100.0)
Wages 5.4 (100.0) 4.5 (100.0) 5.0 (100.0)
Capital 10.9 (100.0) 11.5 (100.0) 11.2 (100.0)
Exports 14.8 (100.0) 4.6 (100.0) 10.4 (100.0)
Imports 12.0 (100.0) 4.4 (100.0) 8.7 (100.0)
Embodied Technology import 6.3 (100.0) ‐0.3 (100.0) 3.5 (100.0)
Disembodied Technology import 10.6 (100.0) 13.0 (100.0) 11.6 (100.0)
Import of Technology 9.2 (100.0) 1.9 (100.0) 6.1 (100.0)
R&D Expenditure 16.1 (100.0) 12.0 (100.0) 14.3 (100.0)
Note: Figures adjacent to the parenthesis are the average annual growth rates (%). The figure in
parenthesis is the relative average share in total manufacturing during a particular period (%).
Source: Author calculation from Prowess IQ (CMIE database).
For the aggregate manufacturing sector, all variables witnessed growth deceleration
during the post‐financial crisis period, except capital and disembodied technology
imports. The largest declining trend is observed among machinery equipment imports,
which fell from 6 per cent growth during the first period to negative 0.3 per cent during
the second period. The domestic production, measured in terms of firm’s sales, grew at the
17
rate of 8 per cent per annum during the entire period, which is marginally lower during
the first period (9.6%) relative to the second (6.1%) period. In the case of basic production
inputs, the capital stock registered an impressive double‐digit growth rate of 11 per cent
whereas the growth rates of wages are considerably less (5 per cent per annum) for the
entire period. As expected, the shrink in global economic activity during the post‐financial
crisis period is found to have negatively impacted the manufacturing trade of Indian firms.
For instance, both exports and imports, which registered double‐digit growth rates during
1995–2007 period (15 and 12 per cent, respectively), witnessed a sharp deceleration during
the 2008–2016 period (5 per cent and 4 per cent, respectively). Moreover, this pattern is
visible among technology import, too, especially in the case of embodied technology.
However, for disembodied technology, we did not find growth deceleration for the entire
period. In addition, we find double‐digit growth in domestic R&D expenditure (14 per cent
per annum), which reflects the growing technological effort by the Indian manufacturing
in the recent period.
In terms of manufacturing production structure, the largest production share is accounted
for by the intermediate industrial firms (56%) while the smallest share is by the basic goods
producing firms (6%). However, in terms of growth rates, consumer durables (13%) and
capital goods (10%) producing firms have registered impressive performance during the
entire period of study. Although the intermediate goods producing firms represent the
largest share in manufacturing, the growth in real wages is found to be the lowest (4%)
compared to other use‐based segments. Similar to what we observed at the aggregate level,
there is a perceptible fall in sales volume and wages during the post‐2008 period, except
for the real capital stock. The increased specialisation in intermediate goods production is
also reflected in the international exchange of goods as the majority of firms were found to
be exporting (57%) to the world market. On the other side, the major importing products
were found to be largely consumption‐oriented goods (78%). Relatively, the export growth
rates of all use based producing firms were higher than import growth rates, especially
during the first period of study.
4.2 Technological Characteristics of Indian Manufacturing Firms
For examining the technology profile of the Indian manufacturing firms, we have taken
the following variables, namely technology import in terms of equipment and machinery
purchases from abroad (embodied technology import, ETI) and foreign exchange spent
on royalty, technical fees, etc. (disembodied technology import, DETI) and domestic
R&D expenditure that reflect the internal technological capability of the firm. In the
following sub‐section, we provide descriptive statistical evidence from the data sample.
Table 2 reveals that the major importer of technology from abroad belongs to the
intermediate goods sector (52%), followed by the consumer durable sector (22%) and
capital goods sector (13%). Although all these sectors have recorded double‐digit growth
rates during the first period, the growth rate fell across sectors during the second period.
The disaggregation of technology import into embodied and disembodied technology
18
purchases shows the significant presence of intermediate goods sector and consumer
durable producing sector in the sample. The relative comparison reveals that the preferred
mode of technology acquisition among use‐based firms is the capital goods imports from
abroad (see Figure 1). For instance, in the case of intermediate goods producing firms, the
bulk of foreign technology purchases (91%) is through capital goods imports. However, in
relative terms, the growth of disembodied technology import has been substantial across
the use‐based category of firms during this period (see Table 2).
Figure 1
Composition of Embodied and Disembodied Technology Import:
By Use‐based Classification (1995–2016)
Note: BAG=Basic Goods, CAG=Capital Goods, CDG=Consumer Durable Goods, CNDG=Consumer
Non‐Durable Goods, IMG=Intermediate Goods, MAN=Manufacturing.
Source: Authors calculation based on data collected from Prowess IQ, CMIE.
The global slowdown since the post‐2008 period seems to have adversely affected the
growth in embodied technology purchases by the consumer durables, capital goods and
basic goods sectors. All these segments have reported negative rates of growth during this
period. Overall, the technology import (both embodied technology and disembodied
technology) for the overall manufacturing sector have reported a growth rate of 6 per cent
per annum for the entire period of study. However, there is a large discontinuity in
technology purchases as only 194 firms have reported purchasing technology from abroad
continuously during the entire period of study (see Table A2 in the appendix for details).
Compared to the in‐house domestic technological effort, which is proxied by the firms’
expenditure on Research & Development (R&D), we find that firms devote a larger part of
their expenditure to acquiring technology from abroad (see Figure 2). As can be seen from
Figure 2, the technology intensity variable, measured in terms of firms’ sales revenue,
indicates that the share of technology import in total manufacturing is around 1.6 per cent,
77.1 76.7 74.7
84.1
91.2
84.1
22.9 23.3 25.3
15.9
8.8
15.9
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0
BAG CAG CDG CNDG IMG MAN
Embodied Technology Import Disembodied Technology Import
19
which is relatively much larger than the share of domestic R&D of 0.4 per cent. It is striking
to note that, apart from capital goods and consumer non‐durables, the share of domestic
R&D is less than unity for the rest of the sector. The consumer‐durable industry groups
(3%) undertake the largest spending on the import of technology and as before, the mode
of preference is capital‐embodied technology from abroad. The analysis of the top 10 major
importers of technology during the last five years is found to be Reliance Industries Ltd
(12%), Maruti Suzuki India Ltd (8%), Tata Steel Ltd (3%), J S W Steel Ltd (3%) and Hyundai
Motor India Ltd (3%). These firms belong to the intermediate and consumer durable goods
segment (see Table A3 in the appendix for further details).
Figure 2
Technology Intensity in Indian Manufacturing: Select Indicators (%)
Source: Author’s calculation from Prowess IQ, CMIE.
5. Estimation Results
In this section, the econometric results of the production function estimates of
manufacturing firms discussed in the methodology section 3 are given. The direct
estimation of production function incorporating technology factors as factor inputs is
statistically more appropriate and widely preferred in the empirical studies on firm level
productivity in India16. For the aggregate manufacturing sector, the firm‐level estimates
16 As noted before, alternative strategy is to run a standard production function estimation using
labour and capital inputs and derive productivity residually. The resulting productivity can be
modelled as a function of a range of potential determinants including technology inputs. This
Two‐step procedure, however, often produces inefficient and potentially biased coefficients
(Harris, 2005). The bias arises from the potential correlation between productivity determinant
factors and production factor variables used to derive the residual. Some of the prominent Indian
1.2
1.6
2.7
1.0
1.6 1.6
0.1
1.1
0.5
1.1
0.1
0.4
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Basic goods Capital goods Consumer
durables
Consumer non‐
durables
Intermediate
goods
Manufacturing
Embodied Technology Disembodided Technology Technology Import R&D
20
are given in terms of OLS and LP. As discussed earlier, in the presence of simultaneity
between productivity shock and input choice decision by the firm, the OLS estimates are
expected to produce an upward bias in the labour coefficient and downward bias in the
capital coefficient. Therefore, consistent estimates are derived using LP, which is a semi‐
parametric method that tackles the problem of endogeneity. The summary statistics of the
selected empirical variables are given in Table A4 in the appendix. Compared to the
standard input variables, there is a high degree of dispersion among the technology import
variables in the sample. This is expected as not many firms have sourced technology from
abroad on a consistent basis during the study period.
Table 3
Production Function Estimates for Indian Manufacturing Sector (1995–2016) Dependent Variable: lny
Variables OLS LP
(1) (2) (1) (2)
lnl 0.594***
(0.006)
0.595***
(0.006)
0.312***
(0.009)
0.313***
(0.009)
lnk 0.367 ***
(0.005)
0.363 ***
(0.005)
0.477***
(0.021)
0.477***
(0.021)
lnTI 0.040 ***
(0.002)
0.020***
(0.002)
lneTI 0.026 ***
(0.002)
0.017***
(0.002)
lndeTI 0.041***
(0.003)
0.027 ***
(0.004)
Constant 2.106
(0.020)
2.106
(0.020)
𝑅 0.740 0.876
Observations 63090 63090 63090 63090
Note: (a) For OLS Huber‐White robust standard errors are given in parentheses. In LP, bootstrap
standard errors are given in parentheses. (b) *significant at 0.01 level for a two‐tailed test, **at
0.05 level for a two‐tailed test, ***at 0.1 level for a two‐tailed test. (d) lny= log of output, lnl=log
of labour, lnk=log of capital, lneTII=log of embodied technology import, lndeTI= log of
disembodied technology import, lnTI=log of technology import
In Table 3, the econometric result of the manufacturing sector is given. For both OLS and
LP estimation method, we have used two models. In the first model, the technology inputs
appear separately as embodied form and disembodied form. In the second case,
technology import is taken as the sum of both embodied and disembodied forms of
technology purchases. In all estimated models, the results reveal that the standard
studies that have employed direct estimation of production function incorporating technology
import inputs are Basant and Fikkert (1996), Hasan (2002), Parameswaran (2009), and Rijesh
(2015).
21
production function factors such as labour and capital have the expected positive sign and
are significant at 1 per cent level. Comparing the estimates of capital reveals that the OLS
produces substantially lower coefficient size in all cases. For instance, the capital estimates
using OLS is found to be 0.36 while LP results indicate an improvement of 0.48. This
suggests that the problem of simultaneity has been addressed by the LP estimates.
5.1 Impact of Technology Import on Manufacturing Output
The elasticity of output with respect to technology import variables is found to be positive
and statistically significant in the Indian manufacturing sector (see Table 3). The
production function estimates based on LP reveals that a 1 per cent increase in technology
import is associated with a 0.02 per cent increase in output, on average. The disaggregation
of technology import further confirms similar range for the elasticity coefficient. For
instance, a 1 per cent increase in the purchase of capital goods embodied foreign
technology increases manufacturing output by 0.02 per cent while a similar increase in
disembodied technology from abroad increases output by 0.03 per cent. Although the size
of the coefficient is rather small compared to other major inputs, the statistically significant
positive impact of technology import is rather noteworthy.
Table 4
Production Function Estimates for Indian Manufacturing: By Use‐Base Classification (1995–2016) Dependent Variable: lny
Variables Basic Goods Capital Goods
Intermediate
Goods Consumer Goods
(1) (2) (1) (2) (1) (2) (1) (2)
lnl 0.356 ***
(0.046)
0.355 ***
(0.045)
0.258***
(0.018)
0.254***
(0.020)
0.261***
(0.013)
0.265***
(0.013)
0.384***
(0.015)
0.384***
(0.015)
lnk 0.578 ***
(0.139)
0.575***
(0.157)
0.458***
(0.056)
0.468***
(0.052)
0.510***
(0.040)
0.510***
(0.040)
0.391***
(0.036)
0.389***
(0.035)
lnTI 0.020**
(0.008)
0.031***
(0.004)
0.020***
(0.002)
0.018***
(0.002)
lneTI 0.017**
(0.008)
0.027***
(0.004)
0.014***
(0.002)
0.016***
(0.002)
lndeTI 0.017
(0.013)
0.023***
(0.007)
0.036***
(0.007)
0.028***
(0.007)
Observations 2226 2226 10605 10605 24596 24596 25663 25663
Note: (a) All estimates are carried out using LP algorithm. The bootstrap standard errors are given
in parentheses. (b) *significant at 0.01 level for a two‐tailed test, **at 0.05 level for a two‐tailed
test, ***at 0.1 level for a two‐tailed test. (d) lny= log of output, lnl=log of labour, lnk=log of
capital, lneTI=log of embodied technology import, lndeTI= log of disembodied technology
import, lnTI=log of technology import.
In Table 4, we provide the econometric result for the disaggregate manufacturing based
on use‐based classification. All estimates are based on the LP algorithm. Similar to what
we have observed at the aggregate level, the traditional factor inputs are found to be
22
positive and statistically significant at 5 per cent level across all use‐based classifications.
In all cases, the import of technology is found to have a positive and statistically
significant (at 5 per cent level) impact on manufacturing output, except for basic goods.
In the case of basic goods, both embodied and disembodied technologies report a
positive impact on output but the statistical significance is apparent only in case of the
former. For the intermediate and consumer goods producing segment, the output
elasticity of disembodied technology is found to be marginally higher than the embodied
technology. However, for capital goods, the import of tools and equipment embodying
superior technology has a relatively larger impact on output. Overall, the range of output
elasticity with respect to technology import ranges from 0.01 to 0.04 (see Table 4).
Table 5
Productivity Determinant Model Estimation Results: By Use‐based Classification
Variables
Manufacturing
Sector Basic Goods Capital Goods Intermediate Goods Consumer Goods
(1) (2) (1) (2) (1) (2) (1) (2) (1) (2)
lnETI 0.016 **
(0.002)
0.014**
(0.001)
0.018*
(0.006)
‐0.019**
(0.056)
0.024**
(0.007)
0.013**
(0.003)
0.011**
(0.003)
0.017**
(0.001)
0.011**
(0.0054)
0.018**
(0.002)
lnDETI 0.026**
(0.005)
0.058**
(0.002)
‐0.024
(0.0175)
0.064**
(0.008)
0.028**
(0.009)
0.031**
(0.003)
0.011*
(0.006)
0.028**
(0.003)
0.028**
(0.008)
0.087**
(0.002)
lnRD 0.034**
(0.005)
0.061**
(0.001)
0.057**
(0.0240)
0.087**
(0.008)
0.022*
(0.011)
0.085**
(0.003)
0.028**
(0.008)
0.047**
(0.002)
0.029**
(0.009)
0.052**
(0.002)
lnEXP 0.086**
(0.004)
0.055**
(0.001)
0.020
(0.015)
0.023**
(0.006)
0.062**
(0.011)
0.037**
(0.003)
0.095**
(0.009)
0.086**
(0.002)
0.067**
(0.010)
0.051**
(0.002)
Constant 2.098**
(0.027)
2.162**
(0.005)
2.568**
(0.087)
2.308**
(0.057)
2.254**
(0.063)
2.18**
(0.030)
2.049**
(0.052)
1.950**
(0.026)
2.35**
(0.052)
2.303**
(0.026)
Industry
Dummy Yes Yes Yes Yes Yes
Year
Dummy Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
R2 0.167 0.283 0.300 0.242 0.167
F‐test 122.01 10.59
(0.000)
15.93
(0.000)
17.1
(0.000)
12.11
(0.000)
Wald chi2 20985.47
(0.000)
1509.54
(0.000)
5727.55
(0.000)
31237.95
(0.000)
8909.12
(0.000)
Observations 63090 2226 10605 24596 25663
Groups 4616 144 793 1746 1933
Note: (1) The estimated equation (1) is based on Fixed Effect estimator. The equation (2) is based on
Flexible Generalised Least Square method of estimation. (2)Huber‐White robust standard errors
are given in parentheses. (3) **at 0.05 level for a two‐tailed test, *significant at 0.1 level for a two‐
tailed test. (4) lnETI=log of embodied technology import, lnDETI=log of disembodied technology
import, lnRD=log of real Research and Development expenditure, lnEXP=log of real exports.
23
In Table 5, the result of productivity determinant model, which is explained in sub‐section
3.1(b), is given. In this specification, the productivity residual, which is obtained in the first
stage of estimation of a standard Cobb‐Douglas production function, is regressed on
technology import, domestic technological capability (proxied by real expenditure on
R&D) and exports. We have used Fixed Effects (FE) estimator (within estimator) that
control for unobserved firm‐specific heterogeneity that is assumed to be constant over
time. The FE model assumes that the firm‐specific effects are correlated with the
explanatory variables. We use the Huber‐White robust standard errors to control potential
heteroskedasticity in the sample disturbance term. In order to control for certain degree of
correlation between the residuals in the model, we also use Flexible Generalised Least
Square (FGLS) method where the structure of the error covariance matrix is unknown. The
result for the FE is given in column (1) and FGLS is given in column (2).
The alternative specification further confirms our earlier result that reveals a positive and
statistically significant impact of technology import on output.17 The coefficient of
embodied technology and disembodied technology is positive and statistically significant
using FE and FGLS among overall manufacturing as well as use‐based classification,
except basic goods. In the case of basic goods, the FE estimator reports a negative but
statistically insignificant coefficient of disembodied technology while FGLS reports a
negative and statistically significant result for embodied technology. For the rest of the
segments, the coefficients are rather small but positive and highly significant. The range of
ETI coefficient is between ‐0.02 to 0.02 while the coefficient for DETI is between ‐0.02 and
0.09. The low range of coefficient may be because of the lack of active sourcing of
technology from abroad by the firms. Interestingly, the coefficient of domestic
technological capability, which is proxied by R&D expenditure, turned out to be positive
and statistically significant across all categories and estimation techniques. The RD
coefficient ranges from 0.02 to 0.09 across specification. The result points out the significant
role of accumulating technological capability in improving manufacturing productivity.
The export performance indicator, which is taken as a control variable, is found to have a
significant positive impact on productivity across manufacturing.
The econometric results provide considerable evidence to support the conjecture that
since liberalisation, technology import has been an important element in manufacturing
productivity. Moreover, both embodied and disembodied technology serve as an
important factors in determining manufacturing firms output in India. This provides
empirical support to the recent trade and growth theories that argue that increased trade
openness provides opportunities to learn, adopt, and assimilate foreign knowledge in
domestic manufacturing, thus creating dynamic productivity gains over time. By
importing technology, firms are exposed to the technological knowledge of advanced
countries. The use of technology from abroad further enhance and supplement the
technological capability of domestic manufacturing when it enhances indigenous
17 Rijesh (2015) has obtained a similar result for the aggregate as well as technology‐intensive
classification of Indian manufacturing firms during the period of 1995‐2010.
24
technological effort in the long run. The positive impact of R&D on productivity further
substantiates the argument that there are significant gain from investments in R&D. This
will enhance the technological capability and absorptive capacity of the manufacturing
firms in the future. Since technology import is not a sizeable, a pro‐active, systematic
approach will provide further productivity gains in the future.
6. Concluding Remarks
The basic objective of the present study was to examine the impact of foreign technology
import on the productivity of Indian manufacturing firms since liberalisation. The impact
of technology purchases in terms of embodied technology (import of capital goods) and
disembodied technology (purchase of royalties, technical fees, consultation fees etc.) is
considered as a major source of technology diffusion in the emerging catch‐up economies
like India. Since major technological advancement is largely concentrated in a handful of
advanced countries, access to technology is a major factor for the developing economies
that are building their manufacturing capability to compete effectively in the world
market. Since the opening up of the Indian economy, it has become easier for firms to
expand their input choices and adopt wide‐ranging production technology from abroad.
The recent trade and growth theories have long argued that the purchases of technology
through trade often result in substantial productivity (dynamic) gains to the follower
country. In this context, the present study attempted to assess the dynamic benefits of
technology import in terms of productivity improvement in the case of Indian
manufacturing firms for the period 1995–2016.
The impact of technology import on production is examined by fitting an augmented
Cobb‐Douglas production function incorporating both embodied and disembodied forms
of technology inputs from abroad. The direct estimation of production function is carried
out using the semi‐parametric approach proposed by Levinsohn and Petrin (2003), which
controls the potential simultaneity between factor inputs and productivity shocks. The
empirical analysis is based on an unbalanced panel of 4616 manufacturing firms over the
period 1995–2016. The inter‐sectoral dynamics of the relationship is analysed by classifying
the selected sample firms across five use‐based categories, namely basic goods,
intermediate goods, capital goods and consumer durable and non‐durable goods. The
categorisation of firms according to these different uses of production is a novel
contribution to the empirical literature on Indian manufacturing. The firm‐level data is
collected from Prowess IQ (CMIE). For the econometric analysis, we constructed the real
values of output, capital, material, labour and technology import variables using
appropriate price deflators.
The descriptive statistics reveals that the structure of manufacturing is largely dominated
by the consumer goods and intermediate producing goods firms. Most of the industrial
segment has witnessed considerable trade integration, measured in terms of growth in
exports and outputs. However, since the global financial crisis of 2008, manufacturing
25
sector has witnessed considerable growth deceleration, which is reflected in the number of
performance indicators. In terms of technology, only a few firms have imported
technology for the entire study period. However, most of the firms are found to engage in
some mode of technology purchases throughout the study, especially in the form of import
of machinery and equipment embodying better technology. The largest recipient of
technology import is the consumer goods‐producing firms. Relative to the domestic
technological effort, as proxied by the R&D expenses, firms are found to spend much larger
amount on technology from abroad. Although the share of embodied technology import
is quantitatively larger, the disembodied technology import seems to have registered
better growth performance during the study period.
The econometric results provide considerable evidence to substantiate the argument that
technology import has a positive and statistically significant impact on Indian
manufacturing output. The LP estimates, both at the overall manufacturing and at the use‐
based categorisation of firms, do indicate that the relationship between the two is robustly
positive and significant for both embodied as well as disembodied technology import,
except in the case of basic goods producing firms. Although the relative size of the
technology import coefficient is not large (compared to the estimates of labour and capital),
the overwhelming evidence of a positive relationship does indicate that technology import
is a major source of productivity improvement in Indian manufacturing. The result is
further confirmed from an alternative specification where we modelled the determinant of
productivity by taking technology import, R&D and exports. Apart from observing a
significant and positive impact of technology import, the results further draw attention to
the importance of investment in building domestic technological capability through firm
R&D. Overall, the results indicate that, since liberalisation, technology purchased from
abroad has provided opportunities for dynamic productivity gains in Indian
manufacturing.
26
References
Balakrishnan, P., Pushpangadan, K. and Suresh Babu, M. (2000), ‘Trade Liberalisation and
Productivity Growth in Manufacturing: Evidence from Firm‐Level Panel Data’,
Economic and Political Weekly, 35(41), 3679‐82.
Bas, M. (2012). Input‐trade Liberalization and Firm Export Decisions: Evidence from
Argentina. Journal of Development Economics, 97(2), 481–493.
Basant, R., & Fikkert, B. (1996). The effects of R&D, foreign technology purchase, and
domestic and international spillovers on productivity in Indian firms. The Review of
Economics and Statistics, 187‐199.
Cohen, W.M. and D.A. Levinthal, (1990), ʹAbsorptive capacity: A new perspective on
learning and innovation,ʹ Administrative Science Quarterly, 35, 128‐152.
Damijan, J. P., Konings, J., & Polanec, S. (2014), ʹImport Churning and Export Performance
of Multi‐product Firmsʹ, The World Economy, 37(11), 1483‐1506.
De Loecker, J. (2007), ‘Product Differentiation, Multi‐Product Firms and Estimating the
Impact of Trade Liberalization on Productivity’, National Bureau of Economic Research
Working Paper Series 13155.
Goldberg P.K; Khandelwal A.K; Pavcnik N and Topalova P. (2010a), ʹImported
Intermediate Inputs and Domestic Product Growth: Evidence from Indiaʹ, Quarterly
Journal of Economics 125, 1727–1767.
Goldberg, P. K., Khandelwal, A. K., Pavcnik, N., & Topalova, P. (2010b). Multiproduct
firms and product turnover in the developing world: Evidence from India. The Review
of Economics and Statistics, 92(4), 1042‐1049.
Grossman, G.M., Helpman, E., (1991), ʹInnovation and Growth in the Global Economyʹ, The
MIT Press, Cambridge, MA.
Harris, R. (2005). Economics of the Workplace: Special Issue Editorial. Scottish Journal of
Political Economy, 52(3), 323‐343.
Hasan, R. (2002). The impact of imported and domestic technologies on the productivity
of firms: panel data evidence from Indian manufacturing firms. Journal of Development
Economics, 69(1), 23‐49.
Herrerias M.J. and Orts V (2011), ʹImports and growth in China, ʹEconomic Modellingʹ 28,
2811–2819.
Kaldor, N. (1967), ́Strategic Factors in Economic Development,ʹ Ithaca, NY: Cornell University
Press.
Kambhampati, U. (2003), ‘Trade Reforms and the Efficiency of Firms in India’, Oxford
Development Studies, 31(2), 219‐233.
Kathuria, V. (2001), ‘Foreign Firms, Technology Transfer and Knowledge Spillovers To
Indian Manufacturing Firms: A Stochastic Frontier Analysis’, Applied Economics, 33(5),
625‐642.
27
Katrak, H. (1997), ʹDeveloping Countries Imports of Technology, In‐house Technological
Capabilities and Efforts: An Analysis of the Indian Experienceʹ, Journal of Development
Economics, 53, 67‐83.
Kim, Sangho. and Saravanakumar, Muthusamy. (2012), ‘Economic Reform and Total
Factor Productivity Growth in Indian Manufacturing Industries’, Review of Development
Economics, 16(1), 152–166.
Kumar N., (1985), ʹCost of Technology Imports: The Indian Experience,ʹ Economic and
Political Weekly, Vol. 20, No. 35 (Aug. 31), pp. M103‐M109+M112‐M114.
Kumar, N. & Siddharthan, N. S. (1994), ʹTechnology, firm size and export behaviour in
developing countries: the case of Indian enterprises,ʹ The Journal of Development Studies,
31(2), pp. 289–309.
Lall, S. (1992), ́Technological capabilities and industrialization,ʹ World Development 20: 165–
186.
Lall, S. (2001), ʹCompetitiveness, Technology and Skillsʹ, Cheltenham: Edward Elgar.
Lee, J.‐W., (1995), ʹCapital goods imports and long‐run growthʹ, Journal of Development
Economics, 48 (1), 91–110.
Levinsohn, J. and, Petrin, Amil (2003), ‘Estimating Production Functions Using Inputs to
Control for Unobservables’, Review of Economic Studies 70, 317–342.
Marschak, J and W.H Andrews (1944), ‘Random Simultaneous Equations and the Theory
of Production’, Econometrica 12,143–205.
Mazumdar, M; Meenakshi, R. and Ray, S. C. (2009), ‘Output and Input Efficiency of
Manufacturing Firms in India: A Case of the Indian Pharmaceutical Sector’, Working
Paper 219, The Institute for Social and Economic Change, Bangalore.
Mody, A., & Yilmaz, K. (2002), ́ Imported machinery for export competitivenessʹ, The World
Bank Economic Review, 16, 23–48.
Olley, Steven G. and Pakes, Ariel. (1996), ‘The dynamics of productivity in the
telecommunications equipment industry’, Econometrica 64, 1263–1297.
Parameswaran, M. (2009), ‘International Trade, R&D Spillovers and Productivity:
Evidence from Indian Manufacturing Industry,’ Journal of Development Studies, Vol. 45,
No. 8, Pp. 1249–1266.
Quah, D, and J.E. Rauch (1990), ʹOpenness and the Rate of Economic Growth,ʹ Manuscript,
MIT.
Rebelo, S. (1991), Long‐run policy analysis and long‐run growth. Journal of Political
Economy, 99, 500–521.
Rijesh, R. (2015), ‘Technology Import and Manufacturing Productivity in India: Firm Level
Analysis,’ Journal of Industry, Competition and Trade, Vol. 15, No. 4, Pp. 411–434.
Rijesh, R. (2016). Productivity Growth in Organised Manufacturing Sector in India:
Evidence from Technology Intensive Classification of Industries. Artha Vijnana, Vol.
LVIII(2), pp. 121‐148.
28
Rivera‐Batiz, Luis and Paul Romer. (1991), ʹEconomic Integration and Endogenous
Growth.ʹ Quarterly Journal of Economics, 106(2), 531‐555.
Sivadasan, J. (2006), ‘Productivity Consequences of Product Market Liberalisation: Micro‐
evidence from Indian Manufacturing Sector Reforms’, Working Paper No 1062, Ross
School of Business Working Paper Series.
Solow, R. M. (1957), ‘Technical Change and the Aggregate Production Function’, Review of
Economics and Statistics 39(3), 212‐320.
Srivastava, V. (1996), ‘Liberalisation, Productivity and Competition: A Panel Study of Indian
Manufacturing’, Oxford University Press, Delhi.
Subrahmanian, KK and M. Pillai (1976), ʹImplications of Technology Transfer in Export‐
Led Growth Strategyʹ, Economic and Political Weekly, October 30, 1976, pp 1729‐35.
Szirmai, A (2013), ʹManufacturing and Economic Developmentʹ, in Szirmai, A., W. Naudé
and L. Alcorta (eds), Pathways to Industrialisation in the 21st Century, New Challenges and
Emerging Paradigms, Oxford, Oxford University Press.
Thomas R., Narayanan K. (2016) Productivity Heterogeneity and Export Market
Participation: A Study of Indian Manufacturing Firms. In: De Beule F., Narayanan K.
(eds) Globalization of Indian Industries. India Studies in Business and Economics.
Springer, Singapore.
Topalova, Peita and Khandelwal, Amit. (2011), ‘Trade Liberalization and Firm
Productivity: The Case of India’, The Review of Economics and Statistics, MIT Press, vol.
93(3), pages 995‐1009, 02.
Van Beveren, I. (2012), ‘Total Factor Productivity Estimation: A Practical Review’, Journal
of Economic Surveys, 26, 98–128.
Yatchew, A. (2003), Semiparametric Regression for the Applied Econometrician, Cambridge
University press, Cambridge, UK.
Zanello, G., Fu, X., Mohnen, P. and Ventresca, M. (2016), The Creation and Diffusion of
Innovation in Developing Countries: A Systematic Literature Review. Journal of
Economic Surveys, 30: 884–912.
29
Appendix
Table A1
Distribution of Sample firms Selected: By 2‐digit NIC
2‐digit Description Number of firms
10 Food Products 284
11 Beverages 98
12 Tobacco Products 14
13 Textiles 526
14 Wearing Apparel 125
15 Leather and Related Products 56
16 Wood and of Products of Wood and Cork 20
17 Paper and Paper Products 141
18 Printing and Reproduction of Recorded Media 6
19 Coke And Refined Petroleum Products 39
20 Chemicals And Chemical Products 556
21 Pharmaceuticals, Medicinal Chemical and Botanical Products 297
22 Rubber And Plastics Products 330
23 Other Non‐Metallic Mineral Products 191
24 Basic Metals 435
25 Fabricated Metal Products, Except Machinery and Equipment 163
26 computer, Electronic and Optical products 185
27 Electrical Equipment 268
28 Machinery and Equipment n.e.c. 365
29 Motor Vehicles, Trailers and Semi‐Trailers 349
30 Other Transport Equipment 48
31+32 Furniture+ Other Manufacturing 120
Manufacturing Sector 4616
Source: Author compilation from Prowess IQ database, CMIE
30
Table A2
Firms with consistent Sales and Technology Import for the entire period:
By Use‐Based Classification
Use Based
Classification
Number of Firms
Sales Technology
Import
Embodied
Technology
Import
Disembodied
Technology
Import
Basic goods 54 12 9 4
Capital goods 163 39 28 20
Consumer durables 169 48 32 16
Consumer non‐durables 202 30 24 5
Intermediate goods 417 65 52 11
Manufacturing Sector 1005 194 145 56
Source: Authors calculation from Prowess IQ, CMIE
Table A3
Technology Import by Top 10 firms
Company Name NIC Use base classification Technology Import
(₹ Billion) %
Reliance Industries Ltd. 19209 Intermediate goods 306.0 (12.1)
Maruti Suzuki India Ltd. 29101 Consumer durables 200.4 (8.0)
Tata Steel Ltd. 24105 Intermediate goods 83.6 (3.3)
J S W Steel Ltd. 24105 Intermediate goods 81.4 (3.2)
Hyundai Motor India Ltd. 29101 Consumer durables 71.1 (2.8)
Steel Authority Of India Ltd. 24105 Intermediate goods 52.3 (2.1)
Bhushan Steel Ltd. 24105 Intermediate goods 41.8 (1.7)
Hindalco Industries Ltd. 24201 Intermediate goods 40.3 (1.6)
Jindal Steel & Power Ltd. 24102 Intermediate goods 38.1 (1.5)
Indian Oil Corpn. Ltd. 19209 Intermediate goods 34.0 (1.3)
Note: The figures in brackets represent share of technology import by each firm in total technology
import purchases (both embodied and disembodied technology import).
Source: Authors calculation from Prowess IQ, CMIE
31
Table A4
Summary Statistics
Sectors Variables Obs Mean Std. Dev. Min Max Manufacturing
lny 63090 6.598 1.893 ‐2.780 14.914
lnl 63090 3.673 1.764 ‐3.199 11.041
lnk 63090 6.191 1.712 ‐2.547 14.557
lneTI 63090 1.053 1.877 ‐3.328 10.816
LndeTI 63090 0.317 1.095 ‐2.530 10.251
LnTI 63090 1.235 1.980 ‐3.007 11.138
Basic Goods
Lny 2226 7.747 1.912 ‐1.753 12.482
Lnl 2226 4.658 1.793 ‐2.948 8.614
Lnk 2226 7.489 1.972 ‐0.921 12.451
lneTI 2226 1.311 2.080 ‐3.319 8.420
LndeTI 2226 0.536 1.409 ‐2.469 7.920
LnTI 2226 1.628 2.203 ‐2.968 8.530
Intermediate
goods
Lny 24596 6.603 1.832 ‐2.780 14.914
Lnl 24596 3.521 1.717 ‐3.199 11.041
lnk 24596 6.369 1.708 ‐2.127 14.557
lneTI 24596 1.080 1.937 ‐3.328 10.816
lndeTI 24596 0.216 0.913 ‐2.485 9.094
lnTI 24596 1.211 2.003 ‐3.007 11.138
Capital Goods
lny 10605 6.683 1.801 ‐2.468 13.036
lnl 10605 4.008 1.705 ‐3.041 10.622
lnk 10605 6.043 1.591 ‐2.413 12.598
lneTI 10605 1.133 1.868 ‐3.328 9.117
LndeTI 10605 0.564 1.374 ‐2.530 8.583
LnTI 10605 1.424 2.016 ‐3.007 9.361
Consumer Goods
Lny 25663 6.459 1.949 ‐2.636 13.052
Lnl 25663 3.594 1.785 ‐3.199 9.180
Lnk 25663 5.969 1.670 ‐2.547 12.238
lneTI 25663 0.971 1.797 ‐3.328 10.289
lndeTI 25663 0.293 1.077 ‐2.530 10.251
lnTI 25663 1.146 1.911 ‐3.007 10.125
lny= log of output, lnl=log of labour, lnk=log of capital, lneTI=log of embodied technology import,
lndeTI= log of disembodied technology import, lnTI=log of technology import.
32
List of ISID Working Papers
202 Economic Reforms and Indian Pharmaceutical Industry: Implications for Industrial
Upgrading, Dinesh Kumar Abrol, Rollins John & Amitava Guha, March 2018
201 Dimensions of NPAs in Indian Scheduled Commercial Banks, Santosh Kumar Das &
Pradyuman Singh Rawat, March 2018
200 Dimensions of NPAs in Indian Scheduled Commercial Banks, Santosh Kumar Das &
Pradyuman Singh Rawat, March 2018
199 Intensity of Use of Land in Urban Residential Areas, Hariharan Ramachandran, August
2017
198 International Trade and Productivity Growth: Evidence from the Organised
Manufacturing Sector in India, R. Rijesh, February 2017
197 Demonetisation: Macroeconomic Implications for Indian Economy, Santosh Kumar
Das & Pradyuman Shankar Rawat, February 2017
196 Health in the Era of Neo‐Liberalism: A Journey from Stateʹs Provisioning to
Financialization to Achieve UHC, Shailender Kumar, December 2016
195 Contractionary Fiscal Policy and Public Investment: An Empirical Analysis of
Emerging Regional Growth Dynamics in India, Santosh Kumar Das, October 2016
194 Bottled Drinking Water Industry in India: An Economic Analysis, Swadhin Mondal,
September 2016
193 An Analysis of Foreign Acquisitions in India’s Manufacturing Sector, Beena Saraswathy,
August 2016
192 Impact of Trade Liberalisation on the Indian Electronics Industry: Some Aspects of the
Industrial Policy Dynamics of Global Value Chain Engagement, Smitha Francis, July 2016
191 Pharmaceuticals, Product Patent and TRIPS Implementation, Dinesh Abrol, Nidhi Singh, et.
al., March 2016
190 FDI, Technology Transfer and Payments for Know‐How: A Case Study of Automobile
Sector, Swati Verma & K.V.K. Ranganathan, March 2016
189 Post‐TRIPS Contribution of Domestic Firms to Pharmaceutical Innovation in India: An
Assessment, Dinesh Abrol & Nidhi Singh, March 2016
188 Impact of Mergers on Competition in the Indian Manufacturing: An Assessment, Beena
Saraswathy, March 2016
187 Trends in Foreign Investment in Healthcare Sector of India, Reji K. Joseph & K.V.K.
Ranganathan, February 2016
186 Industrial Finance in the Era of Financial Liberalisation in India: Exploring Some
Structural Issues, Santosh Kumar Das, December 2015
185 Private Sector in Healthcare Delivery Market in India: Structure, Growth and
Implications, Shailender Kumar, December 2015
* Most of the working papers are downloadable from the institute’s website: http://isidev.nic.in/ or
http://isid.org.in/