Product Market Competition and Productivity
in Indian Manufacturing Industry
Atsushi Kato
School of Business
Aoyama Gakuin University
Acknowledgements: I am grateful to Bishwanath Goldar and Arup Mitra for their
helpful comments and encouragements. I am also grateful to the participants at the
seminar held at Institute of Economic Growth, Delhi and Center for Economic and
Social Studies, Hyderabad.
Atsushi Kato, School of Business, Aoyama Gakuin University, 4-4-25 Shibuya,
Shibuya-ku, Tokyo 150-8366, Japan
TEL 81-3-3409-8111, FAX 81-3-3797-6870, E-mail: [email protected]
0
Product Market Competition and Productivity
in Indian Manufacturing Industry
Abstract
The effects of product market competition and corporate governance on the levels and
the growth rates of productivity of firms in Indian manufacturing sector are examined
for the decade since the economic reform in 1991. This study refines indices that capture
the extent of market competition. Empirical results indicate that the smaller the
market share of a firm, the higher is the productivity growth of the firm. In contrast, the
Herfindahl index does not have any significant effect. As regards the corporate
governance factor, debt asset ratio is not found to have any significant effect on
productivity or growth rate in productivity, indicating insufficient disciplinary powers
by banking sectors.
JEL Classification No.: L25, L53
Key words: Product Market Competition, Productivity Level, Productivity Growth
1
1. Introduction
This study investigates the effects of product market competition and corporate
governance on both the level and the growth of firm productivity in the Indian
manufacturing sector for the decade from 1991-1992 to 2001-2002. An attempt is made
in this study to construct more elaborate measures that capture the extent of product
market competition.
The Government of India embarked on drastic economic reform in 1991 after a phase of
gradual deregulations during the 1980s. The drastic economic reform included
deregulations in industry, as well as changes in trade regime, while financial reform
was relatively slow through the 1990s. Some scholars examined the effect of the
economic reform and found that there was a decline in the growth rate of productivity in
many industries (for example, Srivastava, 2000; Balakrishnan, Pushpangadan and
Basu, 2000; Goldar, 2004). However, their approaches determine whether productivity
increased or decreased during the 1990s. Thus, it is not evident what measures among
the comprehensive reform package affect productivity. Therefore, more detailed
analysis is called for in the evaluation of the economic reform.
In the empirical studies of the effect of market competition, various indices including
n-firm concentration ratio, Herfindahl index, above normal rent, or questionnaire based
results have been used to proxy for the extent of market competition. In the Indian
context, Goldar (1986a, b) applied concentration ratio to analyze its effect on
productivity, but he did not find statistically significant effects. However, in studies for
other countries, positive effects of product market competition on productivity have
been reported. For instance, based on the data of British firms, Nickell (1996), Nickell,
Wadhwami and Wall (1992), and Nickell, Nicolitsas and Dryden (1997) show that
different kinds of indices of product market competition enhance productivity growth.
However, there is a reasonable doubt among scholars that any kind of index at an
aggregate level does not precisely capture competitive pressure faced by a firm in
product markets. This is because, first, at an aggregate industry level, many different
products are included, which are not necessarily substitutes; second, each firm usually
produces a variety of products, which do not necessarily fall into the same industrial
classification.
2
In this paper we attempt to construct a more elaborate index of the competitive
pressures faced by a firm. We begin with the competition variables (Herfindahl index,
import ratio, market share) at each product market level, where products produced by
each firm can be regarded as substitutes. Then we aggregate them for a respective firm.
In this study, without any prior knowledge of any specific functional form, we simply
calculate a weighted sum of market level competition variables, using as weights the
percentage of the sales of each product in the total sales of the firm. We use these
aggregate competition variables, as well as above normal rent, to see the effect of
competition on productivity.
In addition, we also attempt to see the effect of corporate governance factors on
productivity. Unfortunately, for the period of this study the available data on
shareholding patterns are incomplete. Thus, we just test the effects of the threat of
bankruptcy (proxied by debt asset ratio).
In this study we examine the effects of these explanatory variables on both the level and
growth of productivity. Previous studies usually a priori choose one of them as
performance criterion. The level of productivity is partly a result of the past
accumulation (or deterioration) of productivity growth. Whether a certain set of
variables affect either the level or growth of productivity depends on the speed of the
realization and persistence of their effects. Thus in this study we try both measures.
Our main estimation results show that the smaller the market share of a firm is, the
higher the productivity growth of the firm becomes. Moreover, this effect is more
prominent in a less concentrated market. In contrast, Herfindahl index does not have
any significant effect on either the level or the growth rate of productivity. As regards
corporate governance factor, debt asset ratio does not bear a significant relationship
with productivity.
The results indicate that a policy measure that deprives large firms of market power
may be conducive to productivity growth. In such a situation where a firm can occupy a
small market share confronted by many rival firms, it may face more competitive
pressure and the firm’s manager would make greater efforts to enhance the productivity.
In contrast, the threat of bankruptcy as reflected by debt asset ratio does not seem to
have any positive effects on productivity enhancement. This seems to be partly
explained by the fact that in India a great deal of long-term finance is provided by
3
government controlled banks, which have little incentive to monitor borrowing firms.
From these results, further deregulation of financial sector should be considered in
order to give financial institutions incentives to monitor borrowing firms.
The rest of the paper is organized as follows. Section 2 gives an overview of theoretical
and empirical background of this study. Section 3 discusses these issues in the Indian
context. Estimation models are introduced in Section 4, and the construction of
variables and data are explained in Section 5. Section 6 presents our estimation results.
Section 7 concludes the paper with some policy implications based on these results.
2. The Background
In this study we presume that managers’ efforts play a vital role in the improvement of
firm productivity.1 Then we examine whether product market competition and corporate
governance influence the level or the growth rate of productivity, presumably through
managers’ efforts. In this section we give a brief overview of theoretical and empirical
backgrounds of this study.
2-1. Competition
Casual argument broadly asserts that competition, both domestic and international,
may enhance the productivity of firms or industries by motivating the efforts of
corporate managers.
However, the theoretical literature on the effect of market competition does not clearly
accord with this casual argument. One of the reasons for this is that there are various
definitions of competition in the theoretical literature. Intensification of market
competition is captured by, for instance: the change in the mode of competition, from
monopoly to perfect competition (for example, Arrow, 1962) and from cartel to Cournot
to Bertrand competition (for example, Horn, Lang and Lundgren, 1994); increase in the
number of firms (for example, Loury, 1979; Lee and Wilde, 1980; Dasgupta and Stiglitz,
1980a,b; Reinganum, 1982, 1985; Spence, 1984; Martin, 1993; Krishna, 2001); the
number of other managers to whom each manager is compared (for example,
Holmstrom, 1982; Nalebuff and Stiglitz, 1983; Mookerjee, 1984; Hermalin, 1992);
increase in the price elasticity of demand (for example, Willig, 1987); reduction of the
size of demand (for example, Willig, 1987; Hermalin, 1992); decrease in profits (for
4
example, Schmidt, 1997); increase in the substitutability between products (for example,
Aghion, Dewatripont and Rey, 1997); reduction in the gap between a leader and a
follower (for example, Harris and Vickers, 1987); forward movement towards goal by
one of rivals (for example, Harris and Vickers, 1987); the increase in the ratio of
entrepreneurial firms relative to managerial firms (for example, Hart, 1983; Scharfstein,
1988); existence of a potential rival (for example, Gilbert and Newberry, 1982; Nalebuff
and Stiglitz, 1983). Many of these theoretical models are based on principal agent model,
in which the owners of a firm design incentive scheme for the manager, but the
structure of the scheme depends on various factors including market competition. The
models analyze how the optimal incentive scheme changes in response to the change in
the extent of market competition.
The definitions of competition are so diverse and, furthermore, the results rely on
factors incorporated into these models. Not surprisingly, it is not theoretically evident
whether the intensification of market competition would induce managers to exert more
efforts.
In spite of the ambiguity of theoretical predictions, empirical studies have tried to test
the hypothesis that market competition would increase the efforts of managers, and
thus lead to higher performance, including higher productivity. Empirically, the extent
of competition has been captured by several variables; concentration ratio, above
normal rent, market share and questionnaire based results.2 A connection between
concentration and productivity is based on the presumption that as the market gets
concentrated, it is more likely that the firms in the industry would collude, and less
rigorously, that in a more fragmented market with a larger number of firms competition
should be intense. Thus, a higher concentration is expected to be associated with lower
productivity. The manager of a firm registering high above normal rent may feel less
competitive pressure and thus make less effort, leading to lower productivity. If a firm
has a small market share, it may reflect the fact that the firm is at a disadvantageous
position relative to other firms with larger market shares. Thus, the manager of a firm
with a smaller market share may feel more competitive pressures, and make more
efforts to raise productivity.
With regard to the effect of market competition on productivity, several rigorous
empirical studies have been done. Nickell (1996) shows that the increase in the number
of rival firms or the decrease in rents enhanced the productivity growth of British firms.
5
Nickell, Wadhwami and Wall (1992) show that the increase in market share of a firm
reduces its productivity growth. Nickell, Nicolitsas and Dryden (1997) obtained the
result that higher average rents normalized on value added tend to reduce productivity
growth. Green and Mayes (1991) claim that the extent of competition is among the
important variables that significantly explain the difference in efficiency.
Other scholars claim that deregulation, which is expected to intensify competition,
raises productivity of firms. For example, Olley and Pakes (1996) show that in
telecommunications equipment industry in the US, productivity growth accelerated
after the deregulation of the industry. Similarly, the positive impact of deregulation in
transition economies was reported by Li (1997) for China, by Djankov and Hoekman
(2000) for Bulgaria, and by Grosfeld and Tressel (2002) for Poland. As will be mentioned
in the next section, the effects of the Indian economic deregulation since 1991 have also
been explored extensively.
Another important channel through which competition is conducive to productivity
improvement is the turnover of firms in industries. Competition forces firms with low
productivity to exit, while firms with high productivity remain in the market. Or the
latter may expand its market shares, whereas the former may lose its market shares.
The effect of this turnover has been examined by, among others, Dunne, Roberts and
Samuelson (1988), Baily, Hulten and Cambell (1992), Baldwin (1995), Baily and
Gersbach (1995), Bartesman and Doms (2000), Aw, Chen and Roberts (2001). All of
them indicate that the turnover plays a critical role in the increase in the productivity of
industries. In this paper we do not examine this turnover effect, but we focus on the
effect of product market competition promoted by the deregulations in India in the
1990s.
This paper makes improvements on the existing literature in two respects. First, we
make more refined measures of product market competition. In most of the studies, the
index of product market competition is calculated at a rather aggregated level. However,
at those levels, products produced by sample firms are not necessarily substitutes in the
sense that the decrease in the price of a firm’s product reduces the demand faced by
other firms. In this study, we start with the data at product levels, compiled by the
Center for Monitoring Indian Economy (CMIE). CMIE selects hundreds of main
products and gathers data on those product markets, which is compiled in Industry:
Market Sale and Shares. This data book includes the data on each firm’s sales and
6
market share in each product market, as well as the Herfindahl index of the market
among domestic firms and import ratio in the total domestic market sales.
Each firm produces a variety of products and sells them in different markets. When a
manager makes an important strategic decision or makes efforts with regard to each
product, the manager puts different weights on each product, depending on the
importance of each product. For instance, a manager may probably exert more efforts
towards a product which account for 80 per cent of the firm’s sales than towards the one
with only 1 per cent. In this study we simply assume that the manager puts more
weight on the product that accounts for higher percentages of sales in the firm’s total
sales.3 Accordingly, for each firm we construct a weighted sum of competition related
indices in each product market, namely, import ratios, Herfindahl indices, and market
shares. Furthermore, we also make another competitive pressure variable, above
normal rent, that may capture overall competitive pressure, faced by a firm.
Second, we do not a priori determine the choice of the level or the growth of productivity
as performance criterion. The existing studies vary in the choice of dependent variables.
Some use the level of productivity and others use the growth rate of productivity. In this
study we regress both on the same set of explanatory variables, to see which is more
affected by those explanatory variables.
2-2. Corporate Governance
A properly designed structure of corporate governance is generally expected to enhance
the productivity of firms because it disciplines corporate managers, who might
otherwise seek private benefits rather than firms’ or shareholders’ benefits.
In studies on corporate governance, the dependent variables considered in regression
analysis include Tobin’s q, share price, return on equity (ROE) and return on assets
(ROA), rather than productivity, because of their focus on the returns to shareholders.
With respect to those performance variables, the independent variables that have been
tested include: debt/asset ratio or interest payment over net cash flow as the risk of
default (for example, McConnell and Servaes, 1990; Nickell, Wadhwami and Wall, 1992;
Ofek 1993), the pattern of shareholdings (for example, Morck, Shleifer and Vishny,
1988; McConnell and Servaes, 1990; Ofek, 1993; Mehran, 1995; Denis, Denis and Sarin,
1997) and the identity of ownership such as directors, government, financial
institutions, foreigners and so forth (for example, Lichtenberg and Pushner, 1994;
7
Kaplan and Minton, 1994; Kang and Shivdasani, 1995; Nickell, Nicolitsas and Dryden,
1997), board structures (for example, Yermack, 1996; Core, Holthausen and Larcker,
1999), and the threat of takeovers (for example, Lichtenberg and Siegel, 1990; Mitchell
and Lehn, 1990; Martin and McConnell, 1991; Berger and Ofek, 1996).
Among these studies, the followings focus on the effect of corporate governance
variables on productivity. Nickell, Nicolitsas and Dryden (1997) obtain the result that, if
financial institutions are large shareholders, then a firm’s productivity growth would be
higher in the UK. Similarly, Lichtenberg and Pushner (1994), based on Japanese firm
data, demonstrate that financial intermediaries as large shareholders exert a positive
impact, but cross-shareholdings have a negative impact, on productivity of firms.
Nickell, Wadhwami and Wall (1992) show that firms with a higher debt tend to have a
higher growth rate of productivity.
Among the variables examined in the previous studies, we focus on the risk of default
captured by the debt/asset ratio in this study. One of the reasons for this choice is the
non availability of time series data of shareholding patterns during our sample period.
Another reason for the choice of the debt/asset ratio is that in India debt finance,
especially long-term debt, is obtained mainly from government development financial
institutions (for example, IFCI, IDBI, ICICI4, IIBI). Some argue that the personnel of
government-controlled financial institutions are not well motivated to monitor the
companies that they finance (for example, Banerjee, Cole and Duflo 2003). Therefore,
our conjecture is that in India the debt/asset ratio does not affect either the level or the
growth of productivity of firms.
With respect to the effect of the debt/asset ratio, Jensen (1986) theoretically claims that
if there is abundant cash flow, a manager may seek private benefits. Conversely, if a
firm’s debt is high, that is, the risk of default is high, and then a manager cannot use
the free cash flow for private benefits. On the other hand, signaling theory (for example,
Ross, 1977) argues that a manager signals his/her competence by obtaining funds
through debt finance.
Empirically, the disciplinary role of debt with respect to these dependent variables is
reported by McConnel and Servaes (1990, 1995). Ofek (1993) also shows that executives
are more likely to take actions conducive to good performance as debt/asset ratio
increases. And as already mentioned, Nickell, Wadhwami and Wall (1992) show that
8
firms with higher debt tend to have a higher growth rate of productivity.
3. The Indian Context
There is a large literature regarding the productivity of firms or industries in India.
Among them, the determinants of productivity are investigated in many studies. The
issue that attracts most attention is the effect of the economic deregulations since 1991.
The deregulations in industry include the followings. First, a large part of licensing
requirements were abolished except for several industries. Second, the industries
preserved for public sector were reduced. Third, prior scrutiny, by government, of
investment by MRTP5 companies is no longer required. Fourth, ceilings on foreign
direct investment in domestic firms in some industries were relaxed. These measures of
deregulation are supposed to promote the entry of new domestic and foreign firms, and
encourage various firm activities, leading to higher product market competition.
Srivastava (2000), using firm level data, shows that there is a strong evidence of a
decline in productivity growth rates in the 1990s. He also shows that the extent of
productivity growth varies from state to state, and from industry to industry. This
indicates that the effects of deregulation depend on various conditions.6 Balakrishna,
Pushpangadan and Basu (2000) show some evidence of a decline in the growth rate of
total factor productivity after 1991-1992. Das (2003) also confirms that many three digit
industries record lower productivity growth rates in the 1990s compared to the 1980s.
Conversely, Unel (2003) and TATA Services Limited (2003) claim that productivity
growth accelerated after the economic deregulation in 1991. However, Goldar (2004) has
pointed out the shortcomings in their methodologies and, after correcting the faults,
obtained the result that there was a slowdown in productivity growth rates after 1991.
Among individual deregulation measures undertaken since 1991, the effect of
liberalization of trade policy has been most intensively tested. Even before the
deregulation, Goldar (1986a, b) shows that as the extent of import substitution is higher,
the productivity growth is lower. Krishna and Mitra (1998) investigate the impact of
trade liberalization since 1991 and find its positive effect on productivity growth. Chand
and Sen (2002) obtain the result that higher effective rates of protection lowers the
productivity growth. Topalova (2004) shows that the reductions in trade protection lead
9
to higher level and growth of firm productivity. Das (2003) claims that the easing of
import restrictions on capital goods caused a higher productivity growth of the sector,
relative to intermediate and consumer goods sectors. On the other hand, Balakrishnan,
Pushpangadan and Basu (2000), whose paper is already mentioned above, show an
evidence of a decline in the productivity growth in the industries with significant tariff
reduction after 1991-1992.
The determinants of the productivity other than the economic deregulations, which
have been examined in the Indian context, include the followings: agglomeration
economies (Mitra, 1999, 2000), infrastructure (Mitra, Varoudakis and Veganzones,
2002), research and development, technology imports and its spillover (Ferrantino,
1992; Goldar, 1995; Raut, 1995; Basant and Fikkert, 1996; Katrak, 1997; Fikkert and
Hasan, 1998; Kathuria, 2000, 2001; Hasan, 2002; Ray, 2004) and firm size (Patibandla
1998).
However, as regards market competition there is little research in terms of its effects on
productivity, except for Goldar (1986a,b). Although behind the deregulations since 1991
was the government’s intention to promote the growth of firms and industries through
competition, it has not been rigorously tested whether market competition contributed
to productivity. This article attempts to partially fill this gap.
As regards corporate governance, many studies recently focused on it in the Indian
context. Among the variables related to corporate governance, foreign ownership has
attracted much attention in India. This may be because the economic deregulations
since 1991 liberalized the participation by foreign companies and investors in Indian
companies and their presence drastically increased in the 1990s. For instance, Goldar,
Renganathan and Banga (2004) show that, in the 1990s, in Indian engineering industry
foreign firms have higher technical efficiency than domestically owned firms. They also
show that domestically owned firms are catching up with foreign firms due to better
access to imported input and positive spillover from foreign owned firms to domestically
owned firms. Banga (2004) shows that Japanese foreign direct investment has a
positive effect on the TFP growth rate, while US foreign direct investments do not.
Other studies choose different performance criteria as dependent variables. Chhibber
and Majumdar (1999) emphasize that foreign ownership has a significant impact on a
firm’s return on sales and return on assets, only when foreign owners have a significant
10
amount of property rights (in their formulation more than 51% of control). Ramaswamy
and Li (2001) show that if foreign ownership is significant in terms of volume of
ownership and board representation, then managers are less likely to pursue unrelated
diversification. Sarkar and Sarkar (2000) also obtain the result that foreign equity has a
positive effect on company values.
Besides foreign ownership, Sarkar and Sarkar (2000) also show that beyond a certain
level, directors’ shareholdings enhance company values (market to book value ratios),
while institutional investors do not. Lending institutions exert effective monitoring if
they possess a substantial amount of equity in the companies that they finance. De
(2003) investigates the structure and extent of interlocking directorates within large
business groups in India. He shows that interlocking directorates are more prevalent in
larger and less heterogeneous firms, and that directorial interlocks improve return on
assets.
In spite of these many studies relating to the effect of corporate governance factors in
India, the effect of debt/asset ratio on productivity has not been examined.7 Considering
the fact that banks are the main providers of finance to firms in India, it is an important
and interesting question whether they play a disciplinary role in corporate governance.
Moreover, as mentioned before, in the literature of corporate governance, studies focus
on indices other than productivity. Because of the importance of productivity increase
for the level of national well-being, we have to examine whether corporate governance
factors not only positively affect financial indices which is important to investors, but
also productivity of firms.
4. Empirical Formulation
In this study we regress the level and the growth of output (value added) on the same
set of level independent variables including market competition variables and corporate
governance variables, controlling for primary inputs (labor and capital). As will be
explained shortly, the coefficients of the independent variables can be interpreted as
their effects on the level and the growth of productivity.
A level of productivity is partly a result of the past accumulation of growth (or decline)
of productivity. Whether a certain set of variables influences level or growth of
11
productivity is a matter of the speed of the realization of its effects and also their
persistence. If an effect is quickly realized and persistent, then a level of productivity at
a point in time reflects the effect that was realized in the past. For instance, the finding
of a new source of raw material due to a manager’s efforts in one period may quickly
increase productivity in that period and its effects persist for a long time. However, if an
effect is gradually realized and is not persistent, it would be difficult to capture the
effect on the level. For instance, the adoption of a new way of production may take a
long time before it reaches its maximum efficiency, and its effect may deteriorate
through time as technological progress occurs in the economy. Thus, it is not easy to
capture the effects on productivity of the factors that induce managers’ efforts. Rather
than a priori choosing a dependent variable, we will see whether independent variables
affect the level or the growth rate of productivity. More precise formulations are
explained below.
Model 1.
α
β
Assuming a Cobb-Douglass production function, Y = AF ( K , L) = AK L , and taking
natural logarithm of both sides, we obtain,
ln Yit = ln Ait + α ln K it + β ln Lit ,
where i is the firm subscript and t is the time subscript. Yit is output, Kit is capital, Lit is
labor. Ait indicates total factor productivity (TFP) of firm i in period t. It is assumed that
ln Ait = γ + δX it + ε it ; that is, the level of TFP is assumed to be a function of Xit
variables. Inserting this yields,
ln Yit = γ + δX it + α ln K it + β ln Lit + ε it ,
Xit is a vector of variables that may possibly affect the TFP level of a firm, and εit is a
random disturbance term, capturing all other shocks. Firm dummy, θi, and time dummy,
θt, are also included. Thus we estimate,
ln Yit = γ + θ i + θ t + δX it + α ln K it + β ln Lit + ε it .
Although we do not directly regress the level of productivity on independent variables,
by this formulation we can interpret the coefficients of Xit as their effects on the
productivity. A possible problem of correlation between primary inputs and disturbance
12
term is partially mitigated by including firm dummies that account for individual firm
effect, and a set of variables that may affect productivity.
Model 2.
α
β
Assuming again a Cobb-Douglass function, Y = AF ( K , L) = AK L , and taking
natural logs and the derivatives with respect to time, we obtain:
Yˆit = Aˆ it + αKˆ it + βLˆ it ,
where ^ indicates the growth rate of each variable. In Model 2 it is assumed that
Aˆ it = γ + δX it + ε it ; that is, the growth rate of TFP is assumed to be a function of Xit
variables. Inserting this yields,
Yˆit = γ + δX it + αKˆ it + β Lˆ it + ε it ,
Xit is a vector of variables that may possibly affect the TFP growth rate of firm i in
period t and is the same set of variables in Model 1 except for a slight difference in time
period. Again firm dummy, θi, and time dummy, θt, are included. Thus, our final
formulation is;
Yˆit = γ + θ i + θ t + δX it + αKˆ it + β Lˆ it + ε it .
Notice again that we examine the effects of a set of level variables on both level and
growth of productivity. The same excuses apply as in Model 1 with respect to correlation
between primary inputs and disturbance terms.
5. Data and variable construction
In this article I focus on the firms operating in the following eight industries: chemicals,
plastics and rubbers, non-metallic mineral products, base metals, non-electrical
machinery, electrical machinery, electronics, and transport equipment. This choice
13
depends on unverified presumption that these industries are comparatively more
technology intensive, than other industries in manufacturing. Thus, the productivity is
expected to be more sensitive to managers’ efforts in response to the change in
competition variables and corporate governance variables.
My data set covers the period from 1991-1992 through 2001-20028. Since drastic
deregulations of industries began in 1991, this decade should be appropriate to analyze
the effect of market competition caused by the deregulations since 1991. The primary
source of data is PROWESS, compiled by Center for Monitoring Indian Economy(CMIE).
Financial data is obtained from this data source. Other data sources are mentioned
below.
Competition variables
The variables that capture the extent of market competition (henceforth referred to as
competition variables, or CV) are obtained from Industry: Market Size and Shares,
published annually by CMIE. This data source includes the data, for each main product,
on (1) the Herfindahl index (referred to as HH), (2) the import ratio of the product
market (IMP), and (3) market share of each firm (MS). The values of HH and IMP are
common to all the firms in the same product market, while MS is specific to a firm.
CMIE calculates the Herfindahl index in Industry: Market Size and Shares, based only
on the domestic market sales of domestically operating firms. Namely, import is
excluded in the calculation of HH.
Each firm sells a variety of products, and the shares of each product’s sales in the firm’s
total sales vary. In addition, each product is exposed to different degrees of product
market competition. It is not obvious how managers allocate their efforts towards
various products, as well as how responsive they are to the change in product market
competition. Managers of firms do not necessarily put an equal weight on each product.
For an initial trial, it is reasonable to suppose that the more a product’s sales is, the
more efforts towards the product a manager exerts. Accordingly, in order to calculate
aggregate variables that indicate the extent of competition faced by a firm, we simply
calculate a weighted sum of competition variables, using the percentage of each
product’s sales in the firm’s total sales as weights. Note that these aggregated
competition variables are firm specific because not only each firm produces a different
set of products, but also the share of each product in a firm’s gross sales varies from firm
14
to firm.
A weighted sum of a competition variable of firm i (WSCVi) is:
ni
WSCVi = ∑ sij × CVij
j =1
where CVij = IMPj , HH j , MSij
where j is the index of a product that firm i produces, sij is the product j’s sales share in
firm i’s total sales and ni is the index of a product firm i produces. Note that IMP and
HH are product market specific variables. Thus they are identical across firms in each
product market. A problem is that CMIE’s Industry: Market Size and Shares does not
cover all the products existing in the economy. Therefore, I make the following
normalization on WSCVi, to obtain an aggregate index of competition variables (ACVi).
ACVi =
1
WSCVi ,
∑ sij
j
∑
j
sij is the sum of the sales of products in firm i’s total sales, covered in our data set.
In the followings we simply omit “A” in front of CVs. We only consider the observations
that have
∑
j
sij more than 90 per cent, so that aggregate competition variables do not
deviate from actual competitive pressure so much.
In the following regression we test the interactive effect of different competition
variables. Interactive terms among CVs are calculated in each product market, and
then aggregated in the same manner as respective competition variables.
In addition to these aggregate competition variables, RENT is constructed by earnings
before taxes and interest payments, plus depreciation, minus cost of capital times the
sum of total equity and total borrowings, normalized on total assets. User cost of capital
is proxied by the lending rate of interest9 deflated for inflation plus a constant
depreciation rate (7.1%).10 RENT is expected to capture above normal profit which
should reflect the overall extent of competition faced by a firm.
15
In order to mitigate the simultaneity problem, we use competition variables in the
previous period in the regression of Model 1. In Model 2 we use competition variables in
the current period to see its effects on the TFP growth rate from this period to the next
period.
Corporate governance variables.
To see the effect of financial market discipline on managers, in this study we focus on
debt asset ratio (DA). DA is calculated by dividing total borrowings by total asset. The
debt asset ratio is expected to capture the extent of threat of bankruptcy. When DA is
high, corporate managers not only make a greater effort but also refrain from diverting
cash flow towards unproductive projects. However, there is another theory that predicts
a positive correlation between DA and productivity. The signaling theory implies that a
manager uses debt finance if the manager is confident of future profits. We also test the
same effect by interest payment over cash flow, as well as debt equity ratio. They are
expected to have a positive coefficient sign. On the other hand, high DA may constrain a
firm from further borrowing, which may hamper possible productive investment. If this
effect is dominant, then the coefficient of DA can be negative.
As with competition variables DA in the previous period is used in Model 1, while in
Model 2 DA in the current period is used.
Control variables
To control for other possible effects, firm size (logarithms of total assets) and age of a
firm are included.11 Firm size may affect productivity through various channels
including (dis)economies of scale or scope, and bureaucratic inefficiency. The age of a
firm may possibly capture learning curve effects or exhaustion of new methods of
improving productivity.12
As mentioned above, year dummies are included to control for common shocks to all
firms in a specific year. Firm dummies are included so that other firm specific effects not
captured by the above variables would be controlled.
Note that industry dummies are not included because firms in the data set may operate
in various industries. One of the main features of this study is its construction of an
aggregate measure of competitive pressure for each firm when it operates in multiple
markets.
16
Primary inputs and output
Output
We constructed deflated gross value added as a measure of output. The deflators are
calculated, just as we did to obtain competition variables. Namely, the deflator applied
to value added of firm i is:
nik
DEFi = ∑ sik × DEFk ,
k =1
where k indicates a two digit level industry in which firm i operates and sik is the
percentage share of sales in industry k in firm i’s total sales and nik is the index of
industries in which firm i operates. DEFk is the wholesale price index of industry k.
Labor
PROWESS contains the data on wages and salaries. I divide this number by average
wage of each firm. First, the average wage of each industry is obtained by dividing the
total emoluments of the industry by total mandays employed in the industry, the data of
which are available from Annual Surveys of Industries. Since I do not have data on how
many workers are employed for each product, the average wage of firm i is calculated as
weighted average of each industry’s average wage, with weights being the percentage
sales in each industry in firm i’s total sales.13
Capital
To a large extent the study followed the method used by Srivastava (2000), which is
explained in Appendix. The main differences between Srivastava’s method and that
used in this study are as follows. First, although Srivastava (2000) uses a balanced data
set, we use an unbalanced data set, in which companies differ in the initial year. We
calculate the estimated real capital stock in the initial year of each firm in our data set,
using the method Srivastava took to obtain the capital stock for the base year in his
study. Second, although Srivastava (2000) divided the capital into plant and machinery,
buildings and other capital, we did not divide it at this stage.
6. Results
17
6-1. Time Trends of Variables
Table 1 shows the simple correlations among variables. Since we have an unbalanced
panel data set and the software selects a set of observations for each regression, which
have the data for both dependent and independent variables, the set of observations
differ across regressions. Thus, the correlation coefficients in Table 1 may be slightly
different from the correlations that could have been obtained from the set of
observations for each regression.
Among independent variables there does not seem to be any serious problem of
multicolinearity, except for that between HH and MS, which is 0.5504. Accordingly,
these variables are separately regressed on. In Table 1 we include price-cost margin
(PCM) to see its correlation with GVA, because competition may not only enhance
productivity but also curtail profits. Since GVA partly reflect profit rates, these
offsetting effects may blur the productivity enhancing effects. However, there is very
little linear correlation between GVA and PCM. Thus, we could interpret the coefficients
of competition and corporate governance variables as mainly reflecting their effects on
productivity.
Figure 1 represents the time trends of competition variables from 1992-1993 to
2001-2003. We omit the data in 1991-1992 because the data in 1991-1992 is very few. It
is seen that IMP edges up from 5.5 per cent in 1992-1993 to 10 per cent in 2001-2002.
Conversely, HH gradually declines from 18 per cent in 1992-1993 to 13 per cent in
2001-2002. MS of sample firms also went down from 10.5 per cent to 7.5 per cent for the
same period. These changes seem to reflect the effects of economic deregulations since
1991. With respect to RENT, which captures above normal rents, increased from
1992-1993 to 1995-1996 but thereafter remained low, though there were slight bounce
backs in 1998-1999 and 2001-2002.
Figure 2 shows the time trends of DA with RENT. Since DA is a stock variable, whereas
RENT is a flow variable, there does not need to be correlation between DA and RENT.
However, we see reverse movements between them. Until 1995-1996, DA tends to
decline whereas RENT goes up. Since 1996-1997, DA gradually accumulates, while
RENT generally stays low.
18
These trends seem to indicate that throughout the 1990s product market competition
gets intensified, including the competition from imports. Initially, after the beginning of
economic reforms sample firms enjoyed higher rents, probably due to unreaped
economic opportunities. But after a while the rents tend to decline due to competition,
and this may have caused accumulation of debt among sample firms.
We are interested in whether the intensification of product market competition and
accumulation of debt/asset ratio have positive effects on productivity.
6-2. Estimation Results
Tables 2 and 3 show our estimation results. The results of Hausman tests indicate that
fixed effects model is preferred for all the formulations except for a few cases whose
results are not a part of our main results. All the formulations which we are interested
in are in favor of fixed effects model. F tests show that the existence of individual firm
effects is accepted against the hypothesis that constant terms are equal across firms.
The Results of Model 1
As regards the results of Model 1, the debt asset ratio has a statistically significant
negative coefficient, indicating that debt asset ratio is negatively correlated with the
level of productivity.
Among three competition variables, only import ratio, if any, seems to affect the level of
productivity, but the coefficient is negative and significant only at 10 per cent
significance level. This result arises probably because we cannot fully distinguish the
effect of market competition on profits from that on the improvement of productivity. In
the 1990s the regulations on trade were liberalized. Since our dependent variable (gross
value added) could be correlated with profit rate, productivity may appear to have
declined in the markets where higher competition from imports curtails the profits rate.
With respect to control variables, age and size variables are consistently statistically
significant (at least at 10 per cent significance level). The results indicate that the level
of productivity is likely to be high as firms get old, and as firms become large.
Based on these results it is reasonable to infer that market competition variables did
19
not influence the level of productivity in the 1990s, at least when we are concerned with
the short time span. We interpret these results as follows. A large part of the level of
productivity is the persistent effect of accumulated productivity growth. Thus, even if a
manager faces a high competitive pressure in a current period, his or her efforts cannot
realize quick effects on the level of productivity.
Moreover, it seems that above results in Model 1 reflect reverse causalities with respect
to the result of DA. If the productivity level of a firm is consistently low, then the firm
has to depend on debt rather than internal money, and debts accumulate in the long run.
Thus, a stock variable (debt/asset ratio) is negatively correlated with another stock
variable (level of productivity). In a similar vein, if the productivity level of a firm is
high, the firm would live longer and the firm’s size tends to become large. Thus the
effects of reverse causalities may also appear here.
The Results of Model 2
In contrast with Model 1, the debt asset ratios are far from being significant in all
formulations. This result indicates that the debt asset ratio does not restrict the
availability of necessary funds for productive investment. If so, then DA should also
have a negative coefficient in Model 2 as well as in Model 1. On the other hand, it does
not support the hypothesis that DA accelerates the growth of productivity by motivating
managers in fear of default, either.
If we estimate three competition variables separately, then the Hausman tests indicate
that RE models are preferred for the estimation of IMP (in column one) and HH (in
column two) variables. However, in those formulations neither competition variables
are statistically significant (IMP variable is negative and significant only at 10 per cent
significance level for fixed effects model). The R-squared are very low for these random
effects models, too.
For the estimation for the effect of MS in column three, the Hausman test implies that
we should choose FE model and MS’s coefficient is negative and statistically significant
at 1 per cent level. This indicates that a firm with a smaller market share tends to raise
productivity rapidly.
If we combine three variables all together, MS is still negative and significant in column
20
five. In this regression IMP is also negative and significant, indicating that as import
ratio increases, the growth rate of productivity decline. A possible explanation is that
domestic firms have not been able to withstand foreign competition. Thus the increase
in import may have curtailed the growth of profit rates and also the growth of value
added may have been reduced.
When interactive terms among three competition variables are included, the interactive
term between HH and MS is positive and significant in column six. This indicates that
as market concentration increases, the effect of MS on the productivity growth declines.
In other words, in a less concentrated market, a firm with a smaller market share raises
productivity more rapidly. If we could say that competitive pressure is high in a less
concentrated market with each firm occupying a small market share, then our
estimation indicates that product market competition will enhance productivity growth
rates.
It is worthwhile to note that throughout the models HH variable is rarely statistically
significant. The result is consistent with our claim that competitive pressure is firm
specific, therefore the competition variable at an aggregated industry level does not well
capture the extent of competitive pressure faced by each firm. Our result is in accord
with Goldar (1986a, b), who obtains the result that concentration does not significantly
affect the productivity growth.
7. Conclusion
In this paper we examine the effects of product market competition and corporate
governance on the level and the growth rate of TFP. Our findings suggest that a firm
with a smaller market share in a less concentrated market is likely to achieve higher
productivity growth. Another implication of our empirical results is that a concentration
ratio (Herfindahl index in this paper) is not a good proxy for competitive pressure that
corporate managers would feel. Aggregate Herfindahl indices do not have a significant
effect in this study.
Regarding the corporate governance related variable we tested the effect of debt asset
ratio on TFP. It has a negative effect on the level and no effect on the growth rate of TFP.
This result can be partially explained by reverse causality. However, the result is also in
21
contrast to other studies for other countries, where a positive association was reported.
We could rationalize our results in terms of the lack of disciplinary roles under the
government-controlled financial system in India.
Based on our empirical results, it seems that the liberalization of industrial regulations
since 1991 is in the right direction. One concern is that in the 1990s large firms are
becoming larger probably due to the relaxation of restrictions on former MRTP firms.
The government could also reform the financial sector so that the managers of firms will
be more disciplined by financial markets. Further privatization could be one option, but
in terms of productivity, it is more important than privatization itself, to ensure that
managers of firms make a greater effort under the pressure from financial markets.
Appendix
Real Capital Series
With a few differences we follow the method taken by Srivastava (2000). The
replacement cost of the capital stock is calculated as,
Pt i+1 K ti+1 =
Pt i+1
(1 − δ ) Pt i K ti + Pt i+1 I ti+1 ,
Pt i
where δ is the depreciation rate, P is the price of capital, K is capital stock and I is
investment. Although Srivastava divided capital into plant and machinery, buildings
and other capital, we did not do so at this stage.
Implicit deflators for total capital formation are obtained from the National Account
Statistics, in which a series are available for private sector capital formation at current
and base year (1993-94) prices.
To obtain capital at replacement cost for the base year, Srivastava made three
assumptions. We modified his assumptions to our setups as follows.
Assumption 1: No firm has any capital in 1991-1992 of a vintage earlier than 1976-1977.
For firms incorporated after 1977 it is assumed that the earliest vintage of capital in
their capital mix dates to the year of incorporation.
22
Assumption 2: The price of capital has changed at a constant rate at Π=Pt/Pt-1−1 from
15 years before the initial year in the data or the date of incorporation of the firm up to
1991-1992 (or the initial year in the data). Values of Π were obtained by constructing
capital formation price indices from the series for gross fixed capital formation in the
National Accounts Statistics.
Assumption 3
Investment has increased at a constant rate for all firms and the rate of growth of
investment g=It/It-1−1. Here the rate of growth of gross capital formation at 1993-1994
prices is assumed to apply to all firms. Again, different average rates are obtained for
firms established after 1977.
Under these assumptions, if the capital stock of the earliest vintage is τ periods old,
then the revaluation factor for gross fixed assets (RG) is:
R
G
{(1 + g )
=
τ +1
}
− 1 (1 + Π ){(1 + g )(1 + Π ) − 1}
g {(1 + g )(1 + Π )}
τ +1
−1
.
Similarly, the revaluation factor for net fixed assets (RN) is:
R
N
{(1 + g )
=
− (1 − δ )τ +1 }(1 + Π )τ {(1 + g )(1 + Π ) − (1 − d )}(1 − δ )
,
( g + δ ){(1 + g )τ +1 (1 + Π )τ +1 − (1 − d )τ +1 }(1 − d )
τ +1
where d is the rate of accounting depreciation and δ is the economic rate of depreciation.
As in Srivastava (1996) we used 7.1 per cent for δ. We calculate the average accounting
depreciation rate for each firm for d.
We apply these revaluation factors to the initial net fixed asset data in our data set, to
obtain our estimated initial capital stock. After that we obtain the series of capital stock
at replacement cost by the perpetual inventory method, mentioned above.
[Notes]
1. This reasoning is different from other reasoning such as economies of scale and
23
capital utilization. However, these factors could partly be explained by managers’ timely
decisions, which can be included in managers’ efforts.
2. For instance, an answer to a question whether a firm has more than a certain number
of rivals in its product market has been used in empirical studies.
3. Product portfolio analysis, such as growth-share matrix by Boston Consulting Group,
provides a different perspective.
4. ICICI was merged with ICICI Bank in 2002.
5. Monopolies and Restrictive Trade Practices.
6. Mitra (1999) also shows that during the pre-reform period from 1976 to 1992 the
productivity growth rates differ depending on state or industry.
7. Banga and Sinha (2005) investigated the impacts of debt on various variables chosen
by firms and their profitability. But they did not examine its effect on productivity.
8. However, the data in 1991-1992 is very few. Thus we could say that the data covers
the period from 1992-1993 to 2001-2002.
9. The data on the lending rate of interest is obtained from International Financial
Statistics, published by IMF.
10. This is the depreciation rate on capital, which is used by Srivastava (2000).
11. I intend to include R&D expenditure, advertisement expenditure, marketing
expenditure, distribution expenditure. But these variables are far from being significant.
Thus I do not report the results of the estimation including these variables.
Furthermore, these variables are controlled by managers. Thus, they should reflect
market competition and corporate governance.
12. Japan External Trade Organization annually conducts a questionnaire survey on
Japanese firms investing in Southeast and South Asian countries. Last year about 60
per cent of Japanese firms in Singapore answered that very little room for reducing
costs further remains.
13. The data on the number of mandays-employed is missing in ASI for 1997-1998. We
first calculated the average ratios of mandays-employed / employees for the period from
1991-1992 to 1996-1997 for each industry. Then we multiply the data on the number of
employees in 1997-1998 by this average ratio for each industry.
[Reference]
Aghion, P., Dewatripont, M. and Rey, P. (1997) Competition, financial discipline and
growth. Review of Economic Studies, 66, pp. 825-852.
24
Arrow, K. (1962) Economic welfare and the allocation of resource for invention, in: R.R.
Nelson (ed) The Rate and Direction of Inventive Activity: Economic and Social Factors,
(Princeton: Princeton University Press), pp. 609-625.
Aw, B.Y., Chen, X. and Roberts, M.J. (2001) Firm-level evidence on productivity
differentials and turnover in Taiwanese manufacturing. Journal of Development
Economics, 66, pp. 51-86.
Baily, M.N. and Gersbach, H. (1995) Efficiency in manufacturing and the need for global
competition. Brookings Papers on Economic Activity: Microeconomics, pp. 307-347.
Baily, M.N., Hulten, C. and Campbell, D. (1992) Productivity dynamics in
manufacturing plants. Brookings Papers on Economic Activity: Microeconomics, pp.
187-267.
Balakrishnan, P., Pushpangadan, K. and Basu, M.S. (2000) Trade liberalization and
productivity growth in manufacturing: evidence from firm-level panel data. Economic
and Political Weekly, October 7, pp. 3679-3682.
Baldwin, J.R. (1995) The Dynamics of Industrial Competition (Cambridge: Cambridge
University Press).
Banerjee, A.V., Cole, S. and Duflo, E. (2003) Bank financing in India. mimeo.
Banga, R., and Sinha, U.B. (2005) Does the structure of debt affect the output and
investment strategies of the firm?. Journal of Restructuring Finance, 2, pp.1-16.
Bartelsman, E. and Doms, M. (2000) Understanding productivity: lessons from
longitudinal microdata. Journal of Economic Literature, 38, pp. 569-594.
Basant, R. and 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, 78, pp. 187-199.
Berger, P.G. and Ofek, E. (1996) Bustup takeovers of value-destroying diversified firms.
Journal of Finance, 60, pp. 1175-1200.
Chibber, P.K. and Majumdar, S.K. (1999) Foreign ownership and profitability: property
rights, control and the performance of firms in Indian industry. The Journal of Law
and Economics, 42, pp. 209-238.
Chand, S. and Sen, K. (2002) Trade liberalization and productivity growth: evidence
from Indian manufacturing. Review of Development Economics, 6, pp. 120-132.
Core, J.E., Holthausen, R.W. and Larcker, D.F. (1999) Corporate governance, chief
executive officer compensation, and firm performance. Journal of Financial Economics,
51, pp. 371-476.
Das, D.K. (2003) Manufacturing productivity under varying trade regimes: India in the
1980s and 1990s. Working Paper No. 107, Indian Council of Research on International
25
Economic Relations
Dasgupata, P. and Stiglitz, J. (1980a) Industrial structure and the nature of innovative
activity. Economic Journal, 90, pp. 266-293.
Dasgupta, P. and Stiglitz, J. (1980b) Uncertainty, industrial structure and the speed of
R&D. Bell Journal of Economics, 11, pp. 1-28.
De, B. (2003) The incidence and performance effects of interlocking directorates in
emerging market business groups: evidence from India. Working Paper 2003-001,
Indira Gandhi Institute of Development Research
Denis, D.J., Denis, D.K. and Sarin, A. (1997) Ownership structure and top executive
turnover. Journal of Financial Economics, 45, pp. 193-221.
Djankov, S. and Hoekman, B. (2000) Market discipline and corporate efficiency:
evidence from Bulgaria. Canadian Journal of Economics, 33, pp. 190-212.
Dunne, T., Roberts, M.J. and Samuelson, L. (1988) Patterns of firm entry and exit in
U.S. manufacturing industries. Rand Journal of Economics, 19, pp. 495-515.
Ferrantino, M.J. (1992) Technology expenditures, factor intensity, and efficiency in
Indian manufacturing. The Review of Economics and Statistics, 74, pp. 689-700.
Fikkert, B. and Hasan, R. (1998) Returns to scale in a highly regulated economy:
evidence from Indian firms. Journal of Development Economics, 56, pp. 51-79.
Gilbert, R. and Newberry, D. (1982) Preemptive patenting and the persistence of
monopoly. The American Economic Review, 72, pp. 514-526.
Goldar, B. (1986a) Productivity Growth in Indian Industry (New Delhi: Allied
Publishers Private Limited).
Goldar, B. (1986b) Import substitution, industrial concentration and productivity
growth in Indian manufacturing. Oxford Bulletin of Economics and Statistics, 48, pp.
143-164.
Goldar, B. (1995) Technology acquisition and productivity growth: a study of industrial
firms in India. Working Paper E-72-1995, Institute of Economic Growth
Goldar, B. (2004) Productivity trends in Indian manufacturing in the pre- and
post-reform periods. Economic and Political Weekly, Nov.20, pp. 5033-5043.
Goldar, B., Renganthan, V.S. and Banga, R. (2004) Ownership and efficiency in
engineering firms in India, 1990-91 to 1999-2000. Economic and Political Weekly, Jan.
31, 2004, pp. 441-447.
Green, A. and Mayes, D. (1991) Technical inefficiency in manufacturing industries.
Economic Journal, 101, pp. 523-538.
Graham, D.R., Kaplan, D.P. and Sibley, D.S. (1983) Efficiency and competition in the
airline industry. The Bell Journal of Economics, 14, pp. 118-138.
26
Grosfeld, I. and Tressel, T. (2002) Competition and ownership structure: substitutes or
complements? evidence from the Warsaw Stock Exchange. Economics of Transition, 10,
pp. 525-551
Hanazaki, M. and Horiuchi, A. (2000) Is Japan’s financial system efficient?. Oxford
Review of Economic Policy, 16, pp. 61-73.
Harris, C. and Vickers, J. (1987) Racing with uncertainty. Review of Economic Studies,
54, pp. 1-21.
Hart, O. (1983) The market mechanism as an incentive scheme. Bell Journal of
Economics, 14, pp. 366-382.
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, pp. 23-49.
Hermalin, B. (1992) The effects of competition on executive behavior. Rand Journal of
Economics, 23, pp. 350-365.
Holmstrom, B. (1982) Moral hazard in teams. Bell Journal of Economics, 13, pp.
324-341.
Horn, H., Lang, H. and Lundgren, S. (1994) Competition, long run contracts and
internal inefficiencies in firms. European Economic Review, 38, pp. 213-233.
Kang J.K. and Shivdasani, A. (1995) Firm performance, corporate governance, and top
executive turnover in Japan. Journal of Financial Economics, 38, pp. 29-58.
Kaplan, S.N. and Minton, B.A. (1994) Appointments of outsiders to Japanese boards:
determinants and implications for mangers. Journal of Financial Economics, 36, pp.
225-258.
Katrak, H. (1997) Developing countries’ imports of technology, in-house technological
capabilities and efforts: an analysis of the Indian experience. Journal of Development
Economics, 53, pp.67-83.
Kathuria, V. (2000) Productivity spillover from technology transfer to Indian
manufacturing firms. Journal of International Development, 12, pp. 343-369.
Kathuria, V. (2001) Foreign firms, technology transfer and knowledge spillovers to
Indian manufacturing firms: a stochastic frontier analysis. Applied Economics, 33,
pp.625-642.
Krishna, P. (2001) On competition and endogenous firm efficiency. Economic Theory, 18,
pp. 753-760.
Krishna, P. and Mitra, D. (1998) Trade liberalization, market discipline and
productivity growth: new evidence from India. Journal of Development Economics, 56,
pp. 447-462.
27
Lee, T. and Wilde, L. (1980) Market structure and innovation: a reformulation.
Quarterly Journal of Economics, 94, pp. 429-436.
Li, W. (1997) The impact of economic reform on the performance of Chinese state
enterprises, 1980-89. Journal of Political Economy, 105, pp. 1080-1106.
Lichtenberg, F.R. and Pushner, G.M. (1994) Ownership structure and corporate
performance in Japan. Japan and the World Economy, 6, pp. 239-261.
Lichtenberg, F.R. and Siegel, D. (1990) The effects of leveraged buyouts on productivity
and related aspects of firm behavior. Journal of Financial Economics, 27, pp. 165-194.
Lichtenberg, F.R. and Siegel, D. (1991) The impact of R&D investment on productivity:
new evidence using linked R&D-LRD data. Economic Inquiry, 29, pp. 203-229.
Loury, G. (1979) Market structure and innovation. Quarterly Journal of Economics, 93,
pp. 395-410.
Martin, K.J. and McConnell, J.J. (1991) Corporate performance, corporate takeovers,
and management. Journal of Finance, 46, pp. 671-687.
Martin, S. (1993) Endogenous firm efficiency in a Cournot principal-agent model.
Journal of Economic Theory, 59, pp. 445-450.
McConnell, J.J. and Servaes, H. (1990) Additional evidence on equity ownership and
corporate value. Journal of Financial Economics, 27, pp. 595-612.
McConnell, J.J. and Servaes, H. (1995) Equity ownership and two faces of debt. Journal
of Financial Economics, 39, pp. 131-157.
Mehran, H. (1995) Executive compensation structure, ownership, and firm performance.
Journal of Financial Economics, 38, pp. 163-184.
Mitchell, M.L. and Lehn, K. (1990) Do bad bidders become good targets? Journal of
Political Economy, 98, pp. 372-398.
Mitra, A. (1999) Total factor productivity growth and technical efficiency in Indian
industries. Economic and Political Weekly, July 31, pp. 98-105.
Mitra, A. (1999) Agglomeration economies as manifested in technical efficiency at the
firm level. Journal of Urban Economics, 45, pp. 490-500.
Mitra, A. (2000) Total factor productivity growth and urbanization economies: a case of
Indian industries. Review of Urban & Regional Development Studies, 12, pp. 97-108.
Mitra, A., Varoudakis, A. and Vaganzones-Varoudakis, M.A.
(2002) Productivity and
technical efficiency in Indian states’ manufacturing: the role of infrastructure.
Economic Development and Cultural Change, 50, pp. 395-426.
Mookherjee, D. (1984) Optimal incentive schemes with many agents. Review of
Economic Studies, 51, pp. 433-446.
Morck, R., Shleifer, A. and Vishny, R.W. (1988) Management ownership and market
28
valuation: an empirical analysis. Journal of Financial Economics, 20, pp. 293-315.
Nalebuff, B. and Stiglitz, J. (1983) Prizes and incentives: towards a general theory of
compensation and competition. The Bell Journal of Economics, 14, pp. 21-43.
Nickell, S., Wadhwani, S. and Wall, M. (1992) Productivity growth in U.K. companies,
1975-1986. European Economic Review, 36, pp. 1055-1091.
Nickell, S., Nicolitsas, D. and Dryden, N. (1997) What makes firms perform well?
European Economic Review, 41, pp. 783-796.
Nickell, S.J. (1996) Competition and corporate performance. Journal of Political
Economy, 104, pp. 724-746.
Ofek, E. (1993) Capital structure and firm response to poor performance. Journal of
Financial Economics, 34, pp. 3-30.
Olley,
G.S.
and
Pakes,
A.
(1996)
The
dynamics
of
productivity
in
the
telecommunications equipment industry. Econometrica, 64, pp. 1263-1297.
Patibandla, M. (1998) Structure, organizational behavior, and technical efficiency: the
case of an Indian industry. Journal of Economic Behavior and Organization, 34, pp.
419-434.
Raut, L.K. (1995) R&D spillover and productivity growth: evidence from Indian private
firms. Journal of Development Economics, 48, pp. 1-23.
Ray, S. (2004) The changing role of technological factors in explaining efficiency in
Indian firms. Journal of Developing Areas, forthcoming.
Reinganum, J. (1982) A dynamic game of R&D: patent protection and competitive
behavior. Econometrica, 50, pp. 671-688.
Reinganum, J. (1985) Innovation and industry evolution. Quarterly Journal of
Economics, 99, pp. 81-99.
Ross, S. (1977) The determination of financial structure: the incentive signaling
approach. Bell Journal of Economics, 8, pp. 23-40.
Sarkar, J. and Sarkar, S. (2000) Large shareholder activism in corporate governance in
developing countries: evidence from India. International Review of Finance, 1, pp.
161-194.
Sarkar, J. and Sarkar, S. (2004) Corporate governance reforms and corporate sector
development in India, in: R. Darryl and S. Mukherjee (eds) Corporate Governance,
Economic reforms and Development: The Indian Experience, (New Delhi: Oxford
University Press), pp. 166-203.
Scharfstein, D. (1988) Product-market competition and managerial slack. Rand Journal
of Economics, 19, pp. 147-155.
Schmidt, K. (1997) Managerial incentives and product market competition. Review of
29
Economic Studies, 64, pp. 191-213.
Spence,
M.
(1984)
Cost
reduction,
competition
and
industrial
performance.
Econometrica, 52, pp. 101-121.
Srivastava, V. (2000) The impact of India’s economic reforms on industrial productivity,
efficiency and competitiveness: a panel study of Indian companies 1980-1997.
National Council of Applied Economic Research.
Tata Services Limited (2003) Reforms and productivity trends in Indian manufacturing
sector, Department of Economics and Statistics, Tata Services Limited, Mumbai.
Topalova, P. (2004) Trade liberalization and firm productivity: the case of India. Yale
University, available at: www.econ.yale.edu/seminars/NEUDC03/topalova1.pdf.
Unel, B. (2003) Productivity trends in India’s manufacturing sectors in the last two
decades, Working Paper No.03/22, International Monetary Fund.
Vickers, J. (1986) The evolution of market structure when there is a sequence of
innovations. The Journal of Industrial Economics, 35, pp. 1-12.
Willig, R. (1987) Corporate governance and market structure, in: A. Razin and E. Sadka
(eds) Economic Policy in Theory and Practice, (London: Macmillan), pp. 481-494.
Yermack, D. (1996) Higher market valuation of companies with a smaller board of
directors. Journal of Financial Economics, 40, pp. 185-211.
30
Table 1 Covariances among variables
GVA
LABOR
CAPITAL
AGE
SIZE
DA
IMP
HH
MS
RENT
PCM
GVA
1
0.85315
0.85901
0.15019
0.50735
-0.0659
0.008994
0.071216
0.24596
0.086978
0.009971
LABOR CAPITAL
1
0.86721
0.17909
0.46436
-0.04582
-0.00815
0.051996
0.20031
0.039845
0.005094
1
0.10585
0.46151
-0.01626
0.013055
0.046825
0.17779
0.030753
0.00997
AGE
SIZE
DA
1
0.36208
-0.07728
-0.03739
0.032272
0.11409
0.006244
0.008578
1
-0.03752
0.003808
0.12735
0.40134
0.094259
0.025493
1
0.01183
-0.13596
-0.11915
-0.25786
-0.05313
IMP
RENT
PCM
1
-0.23663
1
-0.07928
0.5504
1
-0.02722 0.12164 0.19624
1
-0.00039 -0.00136 0.019415 0.072916
1
Note: GVA=Gross Value Added; DA=Debt/Asset Ratio; IMP=Aggregated Import Ratio;
HH=Aggregated Herfindahl Index; MS=Aggregated Market Shares;
PCM=Price Cost Margin
31
HH
MS
Table 2: Model 1 Regression of logarithm of gross value added
Dependent Variable: Log(Gross Value Added)
Column 1
Column 2
Column 3
Fixed Effects model
Variable
LL
LK
AGE
SIZE
DA(-1)
IMP(-1)
HH(-1)
MS(-1)
RENT(-1)
HH(-1)*MS(-1)
IMP(-1)*MS(-1)
IMP(-1)*HH(-1)
R^2
Adjusted R^2
Coefficient
0.63414
0.35335
0.117483
0.386752
-0.20032
-0.509795
t-statistic
9.0763
8.24287
2.22582
5.96495
-2.94054
-1.70109
***
***
**
***
***
*
0.937378
0.917696
Coefficient
0.639061
0.353069
0.097624
0.386814
-0.210435
t-statistic
9.13755
8.22347
1.89613
5.95849
-3.08451
0.076055
0.295453
***
***
*
***
***
0.937251
0.917529
Coefficient
0.632975
0.351274
0.097851
0.384991
-0.20613
t-statistic
8.91254
8.17387
1.90167
5.92078
-3.02339
0.244744
0.553202
***
***
*
***
***
0.937261
0.917542
Random Effect Model
Variable
LL
LK
AGE
SIZE
DA(-1)
IMP(-1)
HH(-1)
MS(-1)
RENT(-1)
HH(-1)*MS(-1)
IMP(-1)*MS(-1)
IMP(-1)*HH(-1)
R^2
Adjusted R^2
Coefficient
0.613707
0.16041
-4.79E-03
0.320188
-0.205999
-0.096528
***
***
**
***
***
Coefficient
0.608563
0.162075
-4.50E-03
0.314704
-0.211253
0.413845
t-statistic
20.0197
5.87449
-2.41818
8.73887
-5.19991
***
***
**
***
***
Coefficient
0.596911
0.151605
-4.58E-03
0.303345
-0.194921
0.837993
0.836646
0.840148
0.838819
[.0000]
[.0000]
5.0109,
72.874,
FE
1820
1820
t-statistic
19.7921
5.5519
-2.50653
8.46982
-4.83074
4.41845 ***
0.84389
0.842592
[.0000]
[.0000]
4.8338,
72.529,
FE
[.0000]
[.0000]
1820
Values in the parenthses are t-values. *** indicates significance at 1% level, ** 5% level and * 10% level.
LL: Log(Labor); LK:Log(Capital)
Age: The current year minus the year of incorporation
Size: Log(total assets)
DA: Total Borrowings/Total Assets
RENT: Earnings before taxes and interest payments, plus depreciation,
minus the user cost of capital times (total debt + total equity), divided by total assets
IMP: The index of import penetration ratio faced by each firm, as defined in the main text
MS: The index of market share held by each firm, as defined in the main text
HH: The aggregated index of Herfindahl index of each firm, defined in the main text.
32
***
***
**
***
***
2.43301 **
0.991851
F test (P-value)
5.1576,
Hausman test (P-value) 71.406,
FE or RE
FE
No. of observations
t-statistic
20.0161
5.78454
-2.54929
8.84947
-5.04983
-0.512703
Table 2 (Continued)
Dependent Variable: Log(Gross Value Added)
Column 4
Column 5
Coefficient t-statistic
0.535455
9.2442 ***
0.296554
7.83 ***
-4.05E-03 -0.44699
0.427338
7.706 ***
-0.162751
-2.6806 ***
Coefficient
0.549659
0.33054
0.118193
0.419236
-0.299901
-0.344652
0.016823
-0.17704
1.38086
Column 6
Fixed Effects model
Variable
LL
LK
AGE
SIZE
DA(-1)
IMP(-1)
HH(-1)
MS(-1)
RENT(-1)
HH(-1)*MS(-1)
IMP(-1)*MS(-1)
IMP(-1)*HH(-1)
R^2
Adjusted R^2
1.31265
8.23026 ***
0.937349
0.919332
t-statistic
7.81731
7.82953
2.28478
6.57015
-4.37149
-1.15841
0.063808
-0.39193
7.76245
***
***
**
***
***
***
0.940006
0.920978
Coefficient
0.551173
0.330843
0.118926
0.417461
-0.299664
-0.48445
-0.083827
-0.530941
1.38388
0.598609
2.01776
1.28881
t-statistic
7.76671
7.82944
2.28602
6.48865
-4.34965
-1.40099
-0.25994
-0.73022
7.75075
0.508808
0.505171
0.341594
***
***
**
***
***
***
0.94004
0.920851
Random Effect Model
Variable
LL
LK
AGE
SIZE
DA(-1)
IMP(-1)
HH(-1)
MS(-1)
RENT(-1)
HH(-1)*MS(-1)
IMP(-1)*MS(-1)
IMP(-1)*HH(-1)
R^2
Adjusted R^2
Coefficient t-statistic
0.531043
20.9819 ***
0.11742
5.03397 ***
-2.82E-03 -1.83797 *
0.41937
14.236 ***
-0.280686 -7.63168 ***
1.90571
0.864067
0.863072
F test (P-value)
3.9108,
Hausman test (P-value) 142.97,
FE or RE
FE
No. of observations
15.1411 ***
Coefficient
0.519006
0.129151
-3.87E-03
0.386912
-0.361791
0.112959
0.023027
0.641836
2.22608
t-statistic
19.3324
5.23861
-2.48324
11.9297
-9.35369
0.66681
0.134677
2.9392
14.9724
0.872275
0.870998
[.0000]
[.0000]
3.5660,
125.28,
FE
2065
1820
***
***
**
***
***
***
***
Coefficient
0.517085
0.129867
-3.94E-03
0.383648
-0.361307
0.181686
0.154261
1.14426
2.21377
-1.39823
-0.710635
-0.565193
t-statistic
19.2788
5.27016
-2.53422
11.8007
-9.31102
0.82044
0.78713
2.89193
14.8468
-1.59046
-0.26058
-0.20987
0.872862
0.871377
[.0000]
[.0000]
3.5315,
127.77,
FE
[.0000]
[.0000]
1820
Values in the parenthses are t-values. *** indicates significance at 1% level, ** 5% level and * 10% level.
LL: Log(Labor); LK:Log(Capital)
Age: The current year minus the year of incorporation
Size: Log(total assets)
DA: Total Borrowings/Total Assets
RENT: Earnings before taxes and interest payments, plus depreciation,
minus the user cost of capital times (total debt + total equity), divided by total assets
IMP: The index of import penetration ratio faced by each firm, as defined in the main text
MS: The index of market share held by each firm, as defined in the main text
HH: The aggregated index of Herfindahl index of each firm, defined in the main text.
33
***
***
**
***
***
***
***
Table 3: Model 2 Regression of the gross rate of gross value added
Dependent variable: Growth rate of gross value added
Column 1
Column 2
Coefficient t-statistic
0.789159
4.80069 ***
0.017907 0.842017
-0.467066 -1.40592
0.559018
1.62024
0.175937 0.428461
-3.94411 -1.94414 *
Coefficient
0.779004
0.017614
-0.449681
0.409716
0.10991
Column 3
Fixed Effects Model
Variable
DL
DK
SIZE
AGE
DA
IMP
HH
MS
RENT
HH*MS
IMP*MS
IMP*HH
R^2
Adjusted R^2
0.614867
0.331874
0.131523
t-statistic
4.73508 ***
0.827271
-1.35214
1.21731
0.26827
Coefficient
0.762843
0.01712
-0.0615
0.374497
-0.076301
t-statistic
4.70139 ***
0.815299
-0.18453
1.12831
-0.18847
-19.5894
-6.56083 ***
0.333046
0.330265
0.129433
0.348615
0.153284
Random Effects Model
Variable
DL
DK
SIZE
AGE
DA
IMP
HH
MS
RENT
HH*MS
IMP*MS
IMP*HH
R^2
Adjusted R^2
F test (P-value)
Hausman test (P-value)
FE or RE
No. of observations
Coefficient t-statistic
0.858339
6.58103 ***
0.018048 0.936933
0.088116
1.2244
-8.86E-03 -1.35975
-0.05164 -0.41789
-0.483177
-0.5609
Coefficient
0.855887
0.018155
0.087074
-8.62E-03
-0.050177
0.179605
t-statistic
6.56168 ***
0.941972
1.19479
-1.3274
-0.40678
Coefficient
0.856396
0.018042
0.12764
-8.44E-03
-0.056584
0.234506
-0.9841
0.028513
0.021101
1.5621,
13.365,
RE
0.028653
0.021241
[.0000]
[.3431]
1.5501,
9.9268,
RE
1982
1982
t-statistic
6.59595 ***
0.942533
1.54599
-1.25874
-0.45049
-1.06602
0.028411
0.020998
[.0000]
[.6224]
1.6904,
52.688,
FE
[.0000]
[.0000]
1982
Values in the parenthses are t-values. *** indicates significance at 1% level, ** 5% level and * 10% level.
DL: Growth rate of Labor; DK: Growth rate of Capital
Age: The current year minus the year of incorporation
Size: Log(total assets)
DA: Total Borrowings/Total Assets
RENT: Earnings before taxes and interest payments, plus depreciation,
minus the user cost of capital times (total debt + total equity), divided by total assets
IMP: The index of import penetration ratio faced by each firm, as defined in the main text
MS: The index of market share held by each firm, as defined in the main text
HH: The aggregated index of Herfindahl index of each firm, defined in the main text.
34
Table 3 (Continued)
Dependent variable: Growth rate of gross value added
Column 4
Column 5
Column 6
Coefficient t-statistic
0.772563
4.75965 ***
0.017538 0.836442
-0.04375 -0.13121
0.536303
1.57617
0.037697 0.091115
-4.14297
-2.0406 **
2.78402
1.47961
-20.624 -6.68183 ***
-0.37761 -0.45319
Coefficient t-statistic
0.766097
4.72464 ***
0.017176 0.820373
0.094653 0.281087
0.484951
1.42064
-6.12E-03 -0.01479
-4.50773 -1.89126 *
-0.276655 -0.12094
-27.2141 -5.46799 ***
-0.198653 -0.23768
17.7303
2.08921 **
-43.7436 -1.61465
27.143
1.0147
Fixed Effects Model
Variable
DL
DK
SIZE
AGE
DA
IMP
HH
MS
RENT
HH*MS
IMP*MS
IMP*HH
R^2
Adjusted R^2
Coefficient
0.788493
0.017615
-0.439265
0.392538
0.198692
-1.0416
t-statistic
4.79082 ***
0.827732
-1.32107
1.1663
0.477642
-1.25012
0.330903
0.130261
0.351796
0.155758
0.355009
0.158283
Random Effects Model
Variable
DL
DK
SIZE
AGE
DA
IMP
HH
MS
RENT
HH*MS
IMP*MS
IMP*HH
R^2
Adjusted R^2
Coefficient
0.858169
0.01805
0.095805
-8.04E-03
0.050692
-0.803708
-1.25606
0.02933
0.021924
F test (P-value)
1.5515,
Hausman test (P-value) 9.9158,
FE or RE
RE
No. of observations
t-statistic
6.58158 ***
0.936887
1.33374
-1.23477
0.344575
Coefficient t-statistic
0.858819
6.61112 ***
0.01792 0.936704
0.129563
1.55955
-7.78E-03 -1.14827
0.034996 0.232044
-0.530417 -0.58497
0.731181 0.792809
-1.26748 -1.18629
-0.762268 -1.16516
0.028828
0.019923
[.0000]
[.6233]
1.7083,
58.287,
FE
1982
0.02925
0.01885
[.0000]
[.0000]
1982
Coefficient t-statistic
0.859149
6.61406 ***
0.017792 0.930504
0.141201
1.65769 *
-7.76E-03 -1.13921
0.026575 0.175307
0.370875 0.294789
0.837711 0.814844
-1.38948 -0.67848
-0.730419 -1.11306
0.737541 0.178816
-8.46943 -0.55198
-9.71599 -0.67862
1.7274,
65.612,
FE
[.0000]
[.0000]
1982
Values in the parenthses are t-values. *** indicates significance at 1% level, ** 5% level and * 10% level.
DL: Growth rate of Labor; DK: Growth rate of Capital
Age: The current year minus the year of incorporation
Size: Log(total assets)
DA: Total Borrowings/Total Assets
RENT: Earnings before taxes and interest payments, plus depreciation,
minus the user cost of capital times (total debt + total equity), divided by total assets
IMP: The index of import penetration ratio faced by each firm, as defined in the main text
MS: The index of market share held by each firm, as defined in the main text
HH: The aggregated index of Herfindahl index of each firm, defined in the main text.
35
Figure 1 Trends of Competition Variables
0.2
0.15
0.1
IMP
HH
MS
RENT
0.05
0
92-93
93-94
94-95
95-96
96-97
97-98
98-99
99-00
00-01
01-02
-0.05
RENT: Earnings before taxes and interest payments, plus depreciation, minus the user cost of capital times (total debt + total equity), divided by total assets
IMP: The index of import penetration ratio faced by each firm, as defined in the main text
MS: The index of market share held by each firm, as defined in the main text
HH: The aggregated index of Herfindahl index of each firm, as defined in the main text
36
Figure 2 Trends of DA and Rent
0.6
0.5
0.4
0.3
DA
RENT
0.2
0.1
0
92-93
93-94
94-95
95-96
96-97
97-98
98-99
99-00
00-01
01-02
-0.1
DA: Total Borrowings/Total Assets
RENT: Earnings before taxes and interest payments, plus depreciation, minus the user cost of capital times (total debt + total equity), divided by total assets
37