Come
July and, unfailingly, Dataquest-one of two private
agencies that are sources of detailed information
on India's IT sector-releases, over consecutive issues,
its data on performance of the IT industry for the
preceding financial year. Dataquest's information,
unlike that of NASSCOM, covers the whole of the IT
sector, including hardware, software, software services
and IT-enabled services. It also provides detailed
information on the top 20 firms in the composite industry
plus a limited amount of information on the top 200.
That data suggests that during the years since 1991-92,
when the Indian software services industry and, to
an extent, the hardware industry was still in its
infancy, there has been one striking structural feature
characterizing the sector. Over this 17-year period
when industry revenues have grown by more than 150
times or at a compound rate of 34 per cent per annum,
a few firms have routinely dominated the industry.
Thus the share of the top 20 firms in the industry
throughout the period has fluctuated between 47 and
57 per cent, standing at 55 per cent in 1999-2000
and at 56 per cent in 2006-07 (Chart 1). That is concentration
as conventionally measured has been high and relatively
stable. What is more there is evidence that at the
core of the industry concentration is in fact increasing.
According to the results of Dataquest's most recent
survey, the share of the Top 20 firms in the revenues
of the Top 200, which has been increasing consistently
over the last few years, rose sharply from 54 per
cent in 2005-06 to 64 per cent in 2006-07, as compared
to a rise from 50 to 54 per cent between 2004-05 and
2006-07 (Dataquest, July 15, 2007). Acquisitions such
as that of i-Flex by Oracle and a sudden, sharp 136
per cent increase in the revenues of Tech Mahindra
partly explain this trend. But the fact of a high
degree of concentration cannot be denied.
It is indeed true that the Top 20 list has seen the
exit of some firms over the years and the entry of
others and the ranks of those that have remained in
the list for long have changed periodically. But even
assessed in these terms there are a number of firms
that have been represented in the Top 20 league for
a significant number of years or over the period as
a whole. This should be clear from Table 1, which
lists the Top 20 firms in 1999-2000 and 2006-07. During
this period, while industry revenues increased by
close to 7 times at the compound rate of 32 per cent
per annum, well over half the firms remain in the
Top 20 club either in their original avatar or as
entities that have merged with other members in the
list. There appears to be a degree of stability with
respect to industry leaders. Thus, a high degree of
concentration combined with relative stability at
the top seems to be the picture that emerges. And
this is even more significant because Dataquest's
figures relate to the IT ''sector'', comprising of many
industries as conventionally defined.
Chart
1 >>
Table
1 >>
This is surprising given the perception that low barriers
to entry and rapid technological change in the information
technology sector make dominance in terms of either
technology or market share at any given point of time
an inadequate basis for monopoly. If firms have to
remain competitive they have to continuously innovate
and beat the competition, which is extremely difficult
in an industry where technological advance is rapid.
This
view was first developed formally in Massachusetts
Institute of Technology-economist Franklin M. Fisher's
expert testimony in favour of IBM during the thirteen-year
US vs. IBM antitrust battle that began in 1969 and
ended with the case being withdrawn on the grounds
that it was without merit. Fisher and his colleagues
later elaborated IBM's case (Folded, Spindled and
Mutilated: Economic Analysis and U.S. v. IBM, 1983),
which was built on the argument that a company's share
in its designated market at a given point of time
is no indication of either its market power or the
presence of significant barriers to entry into the
industry. That argument rested on grounds that in
a technologically dynamic industry producing a heterogeneous
and differentiated product, a firm with an apparently
large market share could be subject to intense competition,
because it operates ''in a rapidly growing market in
which superior new technologies succeeded each other
with breathtaking speed'' (Carl Kaysen's Foreword).
Further, the information technology industry is one
in which customers were not all small, individually
powerless and poorly informed entities, but large
firms and government agencies that had the knowledge
to assess the appropriateness of prices charged and
the ability to ensure alternative sources of supply
when exploited. History seems to have vindicated this
position held by the defense inasmuch as in time IBM
lost its market leadership in hardware supply and
was finally forced to move out of the hardware business.
The decline in share partly provided the basis for
the withdrawal of the case against IBM 13 years after
it was first registered.
The argument that sheer size and overwhelming market
share need not be the result of anti-competitive practices,
partly rested on the grounds that they could be the
outcome of behaviour, practices and strategies that
reflected 'superior skill, foresight and industry'
rather than the misuse of monopoly. These are practices
that would be inevitable for survival in an intensely
competitive situation, necessitating innovation and
the transfer of the benefits of such innovations to
customers in the form of better products and lower
prices. Only when this was not true could the firm
be seen as adopting anti-competitive practices.
Fisher et. al. identify three flaws in the arguments
of the government and those who held that IBM was
a monopolist, based on evidence that it controlled
around 70 per cent of the market. The first was that
the boundaries drawn to identify the size of the ''market'',
based on which market share was calculated, were arbitrary.
Second, was that they assumed that ''anything'' which
made it expensive or time-consuming to enter an industry
constituted a barrier to entry that facilitated monopolization.
And, third that they held that any price cutting by
an incumbent firm, especially if it was large, was
a predatory practice revealing intent to monopolize.
Implicit in such views, according to them, was the
wrong notion that industries were normally on long-run
equilibrium. Such a notion they held was particularly
wrong in a dynamic industry like the computer industry,
''which has again and again experienced a totally unexpected
explosion of demand as new users of and new uses for
computers came into being. In such an industry, the
competitive process-including the special role of
innovation-is what matters.
The view that it is misuse of monopoly rather than
its presence was the problem was strengthened during
the anti-trust battle between Microsoft and the Federal
and various state governments in the US that began
in 1997. At that time, Microsoft controlled the operating
system running on 90 per cent of personal computers
and reportedly earned a profit of around 40 cents
per revenue dollar (Ken Auletta, Microsoft vs the
U.S. Government, and the Battle o Rule the Digital
Age). Even here the argument was not principally that
Microsoft was using its dominance over the operating
systems (OS) market by exploiting barriers to entry
to overprice its product or slow the pace of technical
advance. Rather, the main issue (which emerged initially
with respect to the Netscape browser), was that Microsoft
was ''leveraging'' monopoly in the OS market, by bundling
new products like Internet Explorer with its operating
system for free and forcing vendors to promote its
browser, while concealing the availability of alternatives.
In the process it was seen as shutting out competition
in new areas to expand its monopoly and reducing consumer
choice. But there was no suggestion that Microsoft
could slow or was slowing the pace of technological
change in its area of monopoly, and yet remaining
dominant.
Overall, therefore, the perception has been and remains
that the information technology sector is one where
barriers to entry are limited and the persistence
of a large market share depends in the final analysis
on sustaining leadership through innovation. However,
this view partly begs the question. The difficulty
with the analysis developed by Fisher et. al. is that
is it rests on three presumptions: (i) that monopoly
does not exist if that structure does not result in
excess pricing and stunted innovation; (ii) that if
a sector is characterized by rapid technological change
it cannot be subject to barriers to entry, which could
either be technological or non-technological in character;
and (iii) monopoly exists when only one firm dominates
the industry for relatively long periods of time.
An industry can be characterized by concentration
for long periods, with different firms accounting
for dominant market share at different points in time.
Barriers to entry need not mean that no firm can break
into the market or expand if it is not the leader,
but that potential threats to incumbent firms can
only come from those which are large in size or have
deep pockets, which in turn makes it possible for
them to buy or develop technologies that can help
them challenge and undermine incumbents. But if there
are, for various reasons, a degree of path dependence
in the capacity to deliver new innovations or bring
them to market successfully, even technology and deep
pockets can favour the incumbent rather than a potential
entrant. Though IBM lost its position of dominance,
it did remain in that position for long. During its
heydays, many of the leading innovations in the computer
industry came from IBM, and not everybody would be
convinced that this was purely because of 'superior
skill, foresight and industry'.
The problem becomes even more complex when these ideas
are applied to an understanding of dominance in the
Indian IT sector. It is easier to apply them to the
hardware segment, where the growing presence and dominance
of international brands in the post-liberalization
period points to the fact that even where new entrants
had the opportunity to grow and develop technological
capabilities in a regime of protection before facing
competition, the global industry leaders can easily
displace them. It obviously applies to the packaged
software industry where few Indian firms have made
a mark, where global leaders dominate the domestic
market, and many of the few indigenously-developed
software ''products'' are losing out in their relevant
markets. But how do we explain persisting dominance
in the software services sector? Here dominance lies
not in the market, which is situated abroad and is
so large that, despite India's scorching pace of growth,
Indian firms still account for an extremely small
market share. The dominance lies in the fact that
among the large number of domestic players catering
to this market, a few (such as TCS, Wipro and Infosys)
account for a large share-much larger than the aggregate
industry figures suggest. It is dominance over supply
rather than dominance in the market that needs to
be explained.
Such dominance cannot be explained by technological
leadership since, as widely accepted, India's presence
is still largely in (technologically) lower-end software
services. This is not an area where technology can
constitute a barrier to entry. The explanation possibly
lies in the ability of leading firms to excel in what
Fisher et. al. refer to as ''other forms of innovation''.
This involves, to start with, ''process innovation'',
or ''a reorganization of the way in which production
is structured'', leading to more efficient ways of
services provision and better global delivery models.
A second form of such innovation is ''the creation
of a management system that keeps decisionmakers in
touch with the marketplace and links that awareness
with the design and manufacturing activities'' as well
as permits quick responses to customer demands and
rapid technological change.
It should be obvious that unlike process innovation
in commodity production, this type of process innovation
in services provision is less transparent and not
easily identified. Yet, it obviously exists and matters,
as suggested by the various forms of certification
that have been in use in the software services industry.
But what is surprising is that the adoption of these
practices, and the process of building an ''image''
or ''brand'' for being a firm which does so, has operated
as a barrier to entry to smaller firms, resulting
in persisting dominance of a few. Even here history,
preexisting market size and deep pockets seem to matter.
