Information Economics and Policy14(2002)253–273

www.elsevier.com/locate/econbase

The software industry and India’s economic development

a b,*

Ashish Arora,Suma Athreye

a Heinz School of Public Policy and Management,Carnegie Mellon University,

5000Forbes Avenue,Pittsburgh,PA15213,USA

b Faculty of Social Sciences,The Open University,Walton Hall,Milton Keynes MK76AA,UK Abstract

This paper assesses the contribution of software to India’s economic development,paying particular attention to the role of software in the absorption of labour and the development of human capital in the economy.The success of the software industry has increased the relative value of professional workers,not only programmers,but also managers and analysts.The growing importance of human capital,in turn,has led to innovative models of entrepreneurship and organization,pioneered by the software sector,and these are slowly taking root and spreading to other sectors of India’s industry.A potentially important and under-appreciated contribution of the software industry is thus its exemplar of good entrepreneurship and corporate governance to the rest of India.Though less visible than the macro contributions to employment and foreign exchange,this role is a source of productivity improvement for all industries,and can have powerful long-term benefits for India’s industrialization and growth.©2002Elsevier Science B.V.All rights reserved. Keywords:Indian software;Software exports;Software and growth;Human capital and development JEL Classification:F1;I2;L8;J5;O0

1.Introduction

In little over a decade,India has emerged as a major exporter of software in the international economy.This remarkable feat has been accomplished through the extraordinary growth of Indian software which,in the last5years,expanded at a compound annual rate of56%.More than two-thirds of this was due to exports,

*Corresponding author.Tel.:144-1908-65-45-56;fax:144-1908-65-44-88.

E-mail addresses:ashish@andrew.cmu.edu(A.Arora),s.s.athreye@open.ac.uk(S.Athreye). 0167-6245/02/$–see front matter©2002Elsevier Science B.V.All rights reserved.

PII:S0167-6245(01)00069-5254A.Arora,S.Athreye/Information Economics and Policy14(2002)253–273 making the industry a major export earner for the country.The proportion of software exports to merchandise exports grew from a negligible amount in1990to over6%in1998–1999.Remarkably,these software exports are largely due to the efforts of domestic rather than foreignfirms.Of the top20exporters in1998–1999,only sixfirms were foreign subsidiaries.

The software industry contributes1%of India’s GNP,but has accounted for over7%of the growth of its GNP(Kumar,2000a,b).In1997,the software industry employed160,000of the total employed workforce of28.245million. Employment in the industry,although constituting only a small fraction of the total,has grown quickly and estimates for the year2000suggest that there will be over410,000IT professionals employed in India.

Can the software sector continue to contribute significantly to economic growth, and what forms are these contributions likely to take?To answer this question,we begin by examining in Section2the factors that have contributed to India’s emerging specialization in software exports.We argue that software services are intensive in human capital and the abundant supply of engineers in the country provides not only an absolute wage advantage,but also a comparative advantage. In Section3we review the factors that constrain the current growth of the software industry on the supply side,in particular the role of underinvestment in literacy and in telecommunications infrastructure.Section4analyses the contribution of software growth to human capital formation.High earnings in software have resulted in considerable private investment in training,and the subsequent emergence of a successful self-financing model of tertiary education in some parts of the country.Section5analyses the impact of software on productivity improvements through the linkage effects of software in the domestic economy. We conclude that this mechanism of productivity improvement is of limited importance in the Indian context.Section5emphasizes the role of high software salaries versus the rest of India’s industry in creating productivity-inducing organizational improvements in software.We also analyse the role of software firms as organizational exemplars.Section6summarizes our main conclusions.

2.Factors favouring the growth of software revenues in India:the role of comparative and absolute advantages

Software is not just another industry.The number of companies that produce software or employ software developers is much greater than the set offirms commonly thought of as softwarefirms,such as Microsoft or Oracle.Indeed,large banks,insurance companies,finance companies,and virtually every organization above a certain size all develop a great deal of software.Much of this software is either developed for a particular user,or consists of a standard‘platform’such as a SAP ERP system or an Oracle accounting system,and is customized to the needs of the user organization.Once in place,these systems have to be maintained andA.Arora,S.Athreye/Information Economics and Policy14(2002)253–273255 enhanced.Some observers claim that over two-thirds of all software development efforts are spent in maintaining and enhancing existing software codes,rather than producing new software(Raymond,1999).

Despite the steady growth of software technology and tools,software develop-ment is still labour intensive and requires relatively little capital.Estimates by Lakha(1994)suggest that labour costs accounted for about70%of all software

1

costs in the early1990s.With the information technology revolution taking hold in the1990s,the demand for IT workers in the developed world has steadily surpassed supply.However,a fairly substantial fraction of these activities can be outsourced and are increasingly conducted away from the user organization.This outsourcing demand has formed the basis of the initial growth of the software industry in India.

The needs of software production seem particularly suited to the resource endowments of the Indian economy.Moreover,scale economies are not a significant barrier to entry.Afirm can—and many do—start off as little more than one software development team.Others have started as temporary employment agencies,requiring a few rooms in which to set up a handful of PCs and a telephone.Further,the production of software is not heavily dependent on physical infrastructure such as roads and ports,although a steady supply of electrical power is critical,as is ready access to PCs,workstations and communication,airports, phones,faxes and increasingly,the Internet.

The initial growth of the software service industry was facilitated by the enlightened‘hands off’policies of the government of India.By the late1980s and early1990s,PC prices had fallen steadily,as had the prices of other equipment. The government allowed liberal imports of both hardware and software tools,and firms were able to provide their own electrical power through a variety of sources, including self-generation.As Table1shows,this growth of software revenue came

2

disproportionately from exports,and thus it is worth exploring the nature of India’s advantage in software exports.

Tables2and3show the extent of the absolute(labour)cost advantage.

1With the decrease in hardware prices and the increase in the wages of software professionals,this estimate is likely to be on the low side for the late1990s.Furthermore,the cost of software is the dominant budgetary item in setting up computerized systems in the west.As the process of computerization accelerates in the world economy,the demand for software will continue to increase. 2As many early observers(Heeks,1996;D’Costa,1998)have noted,this initial growth was markedly dependent on export demand,was based on relatively unsophisticated services,and often was little more than the provision of temporary workers to overseas customers.Arora et al.(2000a)also found that for the most part,Indian softwarefirms were generalists,specializing in terms of neither technology nor vertical industry domains,but competing largely on cost with relatively little to differentiate one from the other.One implication of this is that thefirms surrender the lion’s share of the rents to customers so the net benefits of the mushrooming software industry and its growing productivity are largely passed on to customers,prominently the US which accounts for over60%of India’s total exports of software.256A.Arora,S.Athreye/Information Economics and Policy14(2002)253–273

Table1

Growth of software revenues($million)

Year Export revenues Domestic revenues

198422–

198526–

198638–

198754–

19/90105–

1994/95485350

1995/96734490

1996/971085670

1997/981750950

1998/9926501250

Source:Lakha(1994)forfigures up to19/90;Kumar(2000a,b)for all other years. Amongst developed countries,only Greece shows similar levels in the salaries of software professionals.If one concentrates on the availability of scientists and engineers(Table3),India has one of the largest reserves of these professionals in the world,almost all of whom speak English.

Table3also highlights two other factors.Firstly,India has the potential of expanding this reserve with appropriate investments in primary and secondary education.Secondly,countries such as China and Russia have an even greater supply of trained scientists and engineers.If they were to train a proportion of their scientists in English,these countries could participate more fully in the internation-Table2

International differences in salaries paid to software professionals,1995(US$)

Switzerland USA Canada UK Ireland Greece India Project leader74,00054,00039,00039,00043,00024,00023,000 Business analyst74,00038,00036,00037,00036,00028,00021,000 Systems analyst74,00048,00032,00034,00036,00015,00014,000 Systems designer67,00055,00036,00034,00031,00015,00011,000 Development programmer56,00041,00029,00029,00021,00013,0008000 Support programmer56,00037,00026,00025,00021,00015,0008000 Network analyst/designer67,00049,00032,00031,00026,00015,00014,000 Quality assurance specialist71,00050,00028,00033,00029,00015,00014,000 Database data analyst67,00050,00032,00022,00029,00024,00017,000 Metrics/process specialist74,00048,00029,00031,000–15,00017,000 Documentation/training staff59,00036,00026,00021,000–15,0008000 Test engineer59,00047,00025,00024,000–13,0008000 Source:www.man.ac.uk/idpm/isicost.htm.A.Arora,S.Athreye/Information Economics and Policy14(2002)253–273257 Table3

Reserve of technically trained personnel,selected countries,1995

Country Adult illiteracy Scientists/engineers in Population,Reserve of

rates(%)R&D per million of millions scientists/engineers

population19951995

Males Females1981–1995

USA––3732263981,516

Japan––5677125709,625

Russian Fed.––43581484,984

China102753712004,400 Germany––301682247,312

France––253758147,146

UK––241759142,603

India3562151929140,279

Israel––4826628,956

Vietnam493347324,382

Turkey8282096112,749

Hungary––11571011,570

Greece––774107740

Ireland––187147484 Source:World Bank,World Development Reports(1997)and(1999).

al software industry.But the dynamics of the software industry comes into play as well.As we argue below,the head start enjoyed by the Indian software industry will hurt the prospects of the late comers.

The explanation of the growth of software exports from a country like India because of lower labour costs is well known.It also underlies a somewhat pessimistic outlook for the future of these exports and the software industry.Once the surplus of trained labour is depleted,the cost advantage erodes,making India less attractive in comparison to China and Russia,for instance,as the source for lower value-added services(Heeks,1996).The absence of a sizeable domestic market will compound the problem by depriving the country’s software exporters of the experience needed to ultimately enable them to produce higher value-added services and products(D’Costa,1998).In this respect,the only option is for India to develop a sizeable domestic market and reduce its export dependence. This,however,introduces a dilemma.Perhaps the main reason for the absence of a large and sophisticated domestic market is a relatively unsophisticated economy,which has,until recently,grown at3.5%per year.Thus,the develop-ment of a sizeable domestic software market is likely to be a consequence as much as it is a cause of the growth of this particular industry.Indeed,Arora et al. (2000a)find that conditions between the domestic market and exports market are so diverse that knowledge and experience gained from domestic projects are either258A.Arora,S.Athreye/Information Economics and Policy14(2002)253–273

not applicable,or too costly for overseas customers.Arora et al.conclude that

3,4 domestic market experience is not particularly valuable for the export market. Another way of looking at the growth of software exports is to determine whether India enjoys an underlying comparative advantage in software production `

vis-a-vis the rest of the world.If such a comparative advantage does exist,both the absolute cost advantage and favourable circumstances such as the demand caused by the millennium scare could provide the experience and scale of output needed for dynamic learning processes to kick in and positively influence the growth of the sector’s productivity.In this scenario,the future of Indian software exports is not entirely bleak and increasing productivity could compensate for the erosion of the labour cost advantage.

Productivity levels measured as revenue per employee are lower in India than in other parts of the world(notably Ireland and especially Israel).More importantly, however,compared to other countries,software in India is far more productive than other sectors—the essence of a comparative advantage argument.This is clear when we compare the ratio of labour productivity in software.It is twice that of

5

India’s manufacturing and1.3times that of the US(Table4).The picture is similar for Israel,another country with a fast growing software industry.In an open economy,both India and its trading partners benefit from the country’s specialization in software,and implicitly its imports in less productive economic sectors such as manufacturing.The distribution of these gains is a moot point,of course.Given the country’s heavy reliance on the US for software exports and their undifferentiated nature,it seems likely that productivity improvements in Indian software produce greater benefits for the US rather than India. Although the share of software production in India’s industrial output,exports and employment is increasing,its share in the world market remains low.The picture changes somewhat when we look at only the share of the country’s customized software in the global market:this is estimated to have gone up from 11.9%in1991to18.5%in1999(Kumar,2000b).

Limited infrastructure continues to constrain the industry.The most important of

3This conclusion,however,is conditioned by the types of export projects:simple,small and not very sophisticated.In short,there may be the‘chicken and egg’problem.Given the nature of export projects,sophisticated domestic projects may be of little value,but overseas customers appear unwilling to outsource projects that would enable Indianfirms to acquire the necessary experience. 4A more compelling argument is that the domestic market could be the source for particular types of differentiation,for example,software for multiple languages and using multiple scripts,and for mutual translation.This could be a source of competitive advantage in countries with more than one language where forms,and government and corporate publications have to be in multiple languages.Alter-natively,growth in the level and sophistication of domestic demand mayfinally provide companies with an easy way to‘break’into foreign markets by demonstrating their capabilities.These may yet happen,as the industry matures into more differentiated and distinct segments.

5We use value-added as the index of labour productivity in manufacturing and revenue per employee as the index of labour productivity in software where few material inputs are needed for production.

Table4

Comparative advantage in software production across selected countries,1995

All manufacturing Software Comparative advantage

Revenue per

Output per Value added per Index1Index2

employee(3)

employee(1)employee(2)(3)/(1)(3)/(2)

($’000)

($’000)($’000)

Israel112.2038.30100.000. 2.61 Ireland242.20117.10142.240.59 1.22

India20.80 4.108.930.43 2.18 France205.1377.143161.320.79 2.09 Finland231.9276.1683.460.36 1.10

USA206.0098.20126.020.61 1.28 Sources:authors’computations from the following data sources:data in columns(1)and(2)are taken from the UN Industrial Statistics(1998and1999)published by UNCTAD.Exchange rates used to convert local currencies into dollars are taken from line rf of the International Financial Statistics published by the International Monetary Fund.Data in column(3)are derived from the following national and international sources:India from NASSCOM(www.nasscom.org);Israel from Israeli Association of Software Houses(http://www.iash.org.il);Ireland from National Software Directorate (http://www.nsd.ie),Ireland,and France,Finland and USA from The Software Sector:A Statistical Profile for Selected OECD Countries(OECD http://www.oecd.org/dsti/sti/it/infosoc/index.htm). Figures for Israel are obtained by dividing Israeli software revenues by estimated employment.Figures for Ireland are obtained by excluding multinationals from the calculation,which may therefore, underestimate revenue per employee in software.

these are the availability of power and the quality of the telecommunications infrastructure(bandwidth and,increasingly,limited telephone penetration).In 1996,India had15main telephone lines per1000people,compared with395per 1000for Ireland and446per1000for Israel.The situation is more serious when we consider the penetration of PCs in the total population:1.5computers per1000 people versus145for Ireland and117.6for Israel(WDR1998/99).It is difficult to estimate accurately the extent to which infrastructure constraints have affected productivity.Some indirect evidence,however,is available.Costs for power are the second highest expenditure and many softwarefirms generate their own 6

power.Low bandwidths are also a problem.While current bandwidths are adequate for simple tasks,they could become an obstacle to more complex,higher value-added projects being awarded to the samefirms.For the newly emerging area of e-commerce,the lack of telephone penetration will emerge as an important problem.Here,too,the solutions point to the nature of the problem.Mobile phone penetration(which does not require land lines)experienced the most rapid growth in smaller towns.In this context,the development of mobile telephony and Internet products presents a window of opportunity and growth inasmuch as

6A manager of a leading softwarefirm noted that spending on diesel power generators was the second largest item of thefirm’s capital expenditure budget.Thisfirm claims to have generated4MW in1997,the year in which the interview took place.demand for these is not constrained by telephone lines,or by literacy on the part of mobile phones.

3.Can the Indian success be replicated?Implications for other developing countries

Many researchers have predicted competition from other labour-abundant countries such as China,or even Russia and Ukraine,where economic woes have resulted in large reserves of underemployed engineers and scientists.Countries, including China,are reportedly investing heavily to introduce English language skills to their engineers.Undoubtedly,the current market shares of these countries could be increased—and increased substantially—if abundant skilled labour were the sole determinant of success.But the success of the software industry also reflects a certain level of entrepreneurial and managerial capabilities,as well as the importance of strong links with major markets.In the case of India,these are expatriates working in high-level technical and managerial positions in the west, primarily in the US.

Such links helped Indian entrepreneurs to respond quickly to the growing demand for software services.At a minimum,this meant the ability to recruit programmers,and arrange for and manage outsourcing contracts.As thefirms grew,so did the challenges.Successful enterprises developed capabilities that became the source of India’s competitive advantage.Interviews with US managers reported in Arora et al.(2000a)highlight the importance placed by American companies on the ability of the Indianfirms to mobilize large teams of developers at short notice.This,in turn,places demands on thefirms to develop substantial expertise in recruiting,screening,training and,as discussed earlier,retaining software professionals.As discovered in Russia,this is no trivial matter.Russian firms reportedly complain that getting foreign companies to overcome their hesitation of doing business with the country is a major obstacle to offshore programming.Paucity of quality control and proper management are also handicaps,despite a substantial cost advantage in wages.Russia trails behind India in the number of companies with the ISO9000certification(Santana,2001). Furthermore,Indianfirms are increasing productivity by improving their software development processes,by moving up the value chain,and by developing proprietary development tools(Table5).More recent entrants in the industry have also had some success at developing products.Arora et al.(2000a)find that larger firms(with more than250employees)earn$8000–10,000per employee more than smaller ones.The number of such largefirms has increased over time.Similarly, Arora et al.(2000a)report that Indianfirms rated at CMM level3or higher earn about$6000–10,000per employee more thanfirms without this qualification.AsTable5

India’s manpower and revenues/man-year

Year Manpower Revenue per employee($)

1993–199490,0006198.5

1994–1995118,0006998

1995–1996140,00024.5

1996–1997160,00011,036

1997–1998180,00015,000

1998–1999250,00015,600

Source:Arora et al.(2000a,Table1b).

7

many as32Indian softwarefirms have received the SEI-CMM certification and more than half the companies with the CMM4and5ratings are in India. These ratings demonstrate the significant organizational capability in software development that has been built up over the last decade,making it difficult for

8

other global companies to compete.More importantly,competitors have to contend with higher market visibility and the business connections Indianfirms have been able to establish.Out of Fortune’s500companies,185now outsource their software production to India.Indeed,onsite services have given way to more profitable offshore services with dedicated software centres,an indication of the trust US and Europeanfirms have in the quality of Indian software services. Although not insurmountable,these are formidable barriers for others to overcome, including the country’s own late entrants to the industry.Thus,we are likely to see established Indianfirms leverage their reputation and capability by outsourcing to China and elsewhere,as TCS,Wipro and Infosys are reported to be considering (Sengupta,2001).Similarly,a large Chinese telecomfirm,Huawei Technologies, has set up an R&D centre in Bangalore where180Chinese programmers work alongside the locals.In other words,rather than a zero sum game,China and other nations may be able to participate in the international division of software labour through collaboration with India.

7CMM(Capability Maturity Model)is a structured process for software development associated with the Software Engineering Institute at Carnegie Mellon University.It consists offive‘maturity’levels.Companies or units assessed at level four andfive are capable of controlling,managing and improving software development practices.Though initially developed as a means of providing improved software systems for the Department of Defence in the US,the CMM is becoming popular among Indian software servicefirms as a means of signalling their capability to overseas clients, particularly in the US.

8Arora et al.(2000a)do not,however,find any difference in the productivity of younger and older firms.4.The growth of software and human capital formation:public and private investments in training and the rewards to an engineering

9

education

Although India has a large number of scientists and engineers,it also has one of the lowest literacy rates in the world;52%of the total working population are illiterate.As Table3shows,despite the vast reserve of engineers,their number per million is smaller than in several other countries.There is a corresponding over-reliance on the current reserve of trained but underemployed engineers,for whom the slowly growing and protected economy cannot generate adequate demand.

A large proportion of the employees in softwarefirms are college graduates,as highlighted by a sample survey of nearly60softwarefirms who reported that over 80%of their employees had engineering degrees.Only13%were non-engineers

10

trained in software development.

This preference for engineers was unremarkable and of little consequence at the start of the industry,when demand versus annual supply was small.Currently, over160,000engineers from all disciplines graduate every year.The sharp and sustained growth of the software industry had increased its workforce to nearly 250,000by1998–1999,and estimates suggest that this may reach400,000in the years2000–2001.If the software industry continues to grow at50%per year,then there will be a shortage of engineers,regardless of productivity improvements(see

11

Arora et al.(2001)for more details).

These projections are consistent with other evidence.Wages in the software industry have risen over20%per annum and attrition rates are high.In1998–1999 in a sample of over100enterprises,more than half of thefirms,irrespective of age,size or market orientation,reported the shortage of manpower and employee attrition(Arora et al.,2000a)to be among the three main problems.Virtually all firmsfind it difficult to attract and retain talented software developers despite wages that are substantially above Indian standard.

Public policy has responded with increased investments in engineering colleges, placing greater emphasis on information technology in engineering curricula and on the creation of institutes of information technology(IIIT)similar to the better

9This section draws heavily upon Arora et al.(2001).

10An earlier study(NASSCOM,1999)reported that only2%of all software developers trained in private training institutes join software developmentfirms.

11Recognizing the importance of this fact,many Indian policymakers have called for an‘educational emergency’declaration to ensure that the supply of skilled software developers is increased.Several CEOs of the smaller software developmentfirms and NASSCOM(the professional association representing the views of thesefirms)have begun to argue that the shortage of skilled labour is constraining their ability to grow.(See also Basic Background Report(BR-3)for the National Task Force on Information Technology(IT)and Software Development(SD)submitted to the Prime Minister of India,18th March1999.)known Indian Institutes of Technology.Though superficially reasonable,this is not the answer.These investments are unlikely to have a significant effect on supply in the short run.Moreover,expansion of capacity ignores the problem that the growth of the software industry has tended to siphon off engineering masters and Ph.D students.A recent report noted that the number of engineering Ph.Ds has fallen from675in1987to375in1995.Similarly,the number of engineers with postgraduate training has risen only slowly,from|12,000in1987–19to a little over17,000in1990–1992.Surveys of India’s IITs(the premier technological institutions)show that a very large fraction of postgraduates(90%in some cases) enter the information technology(IT)sector.

Moreover,Table6shows that the majority of India’s engineering capacity is located in just a few states—Maharashtra,Karnataka,Tamil Nadu and Andhra Pradesh,and that the bulk of the capacity is in‘self-financed’colleges,where students receive a much smaller subsidy,if at all,than in the state-financed colleges.An interesting and hitherto unexplored question is the reason why the organizational innovation of self-financing colleges has not diffused to other parts of the country.We can only speculate that this may be related to cultural and political factors as well as the lower returns to investment in human capital in other regions of India.It is no coincidence that the south and west are also economically more advanced.

Even though investments in engineering education are necessary,we believe that a more efficient use of existing human capital resources could provide at least a part of the solution.Implicit in the discussion thus far is the fact that only engineering graduates are adequately suited for the required software tasks.This assumption appears to have shaky foundations.Firstly,the bulk of engineers working in the industry are,in fact,not trained in software engineering,computer science,or related disciplines.Further,a very significant fraction of the work involves developing and refining business applications,databases,and the like. Indeed,initially the majority of the work involved porting applications from one computing platform,typically a mainframe,to another platform such as Unix.This Table6

Number and capacity of engineering colleges in India(approved up to1998–1999),by region Region No.of Sanctioned capacity%of sanctioned capacity

colleges(no.of students)that is self-financed

colleges

Central5094700.52

East2548120.26

North(incl.north-west)14025,4490.42

West14034,1650.74

South(incl.south-west)30882,5970.79

Total663156,4930.69

Source:Ramarao(1998).necessitates certain familiarity with software development tools,but does not

12 require in-depth knowledge of computer architecture or operating systems. Finally,much of the work tends to consist of small projects,with fairly low levels of technical complexity.Arora et al.(2000a)report that the median size of the ‘most important export project’of thefirms in their survey was only150man months,with an average of510.This suggests that the typical export project is even smaller.Moreover,about half of the work was carried out in India,the rest was onsite in the US.

When pressed,most of the managers agreed that they did not specifically require engineers.With proper training,bright graduates from anyfield would suffice. Thus,the preference for engineers in some cases seems to be a quality signal to customers.As one CEO put it:

Take somebody from a good college(any of the top20colleges in India), give him3months of orientation and they are ready to take up a programming assignment.I don’t need all these engineers...But I don’t want to be branded by my customers as a guy who hires NIIT graduates [emphasis added].

(From Sloan Report,Arora et al.,2000b)

This is a clear instance of a‘race to the top’.With limited market power,Indian software exporters try to distinguish themselves from competition by highlighting the quality of their processes and employees,and when possible,their experience. Firms also have quality concerns.Some interviewed managers believe that an engineering education imparts problem-solving skills,methods of logical thinking and learning tools that would promote quick adaptation to changes in technology, domains and tasks.Since Indianfirms provide services across a range of platforms and domains,this is an important asset.Another important consideration is the signalling of qualifications in the labour market.The Indian education system is such that competition for an engineering education is intense,and as a result,an engineering degree is synonymous of such qualities as intelligence and the willingness to work hard.Softwarefirms may value these qualities more than specific substantive engineering knowledge.

13

If this is the case,it is certainly an inefficient allocation of resources.Indeed, the software industry has expanded in part because of its ability to attract engineers from other industries.In our interviews there were a number of instances

12The significant exception is telecommunication-related software,where telecommunication engineers are required.Similarly,companies such as Texas Instruments,Oracle and Microsoft are locating development and R&D centres to tap into the engineering talent.

13A more rational allocation could be the substitution of intelligent and hardworking non-engineers trained in the use of software development tools for engineers wherever possible.Engineers could be used where their substantive knowledge or design ability was of value.

of engineers with highly specialized training(such as VLSI design or satellite systems)working on database design or the development of business application software,and quite a few senior level engineers were drawn from various public sector research and development institutions.Such a transfer of resources reflects the presumed comparative advantage in software development,but there are distortions in the economy,which warrant caution in interpreting the market 14

signals.In addition to changing the composition of economic activity,the increasingly tight market for engineers and managers is also likely to affect the organization of economic activity,such as the balance between capital and labour (or more precisely,between capital and human capital),resulting in organizational innovations.

The clearly increasing payoffs to human capital are also inducing greater investment in human capital.The middle class in India has always relied on education,particularly professional training such as engineering or medicine,as the means for economic advancement.However,with a slowly growing economy, the returns to this investment have not been very high.Precarious publicfinances have limited the ability of central and state governments to expand tertiary education,but the rapid growth of the software sector has created a watershed. One of the most rapidly growing sectors within the software industry is in private training(Table7),where institutions specialize in software development.Sources from NASSCOM(the National Association of Software and Service Companies) estimate that in1998there were3800such trainingfirms,in what was then a$300 million market,although NIIT and Aptech together are believed to have70%of

15

the software training market.Private institutes are also important for upgrading the skills of existing software developers.This option is also used by many engineers to undertake further training in software development on their own.The

Table7

Growth of the information technology training sector in India

Training1996–1997($million)1997–1998($million)Growth(%)

Corporate11.6721.4384

Individual145.24182.4326

Total156.81203.8630

Source:INFAC,Mumbai.

14There might be long-run consequences as well which could damage the indigenous technology development capability.This is similar to the trend,noted earlier,of engineering graduates deserting other disciplines for the IT sector,and even more,of preferring software jobs rather than pursuing postgraduate education.

15Other major vendors include Software Solutions Integrated,LCC Infotech,Tata Infotech,CMC, Indian Institute of Hardware Technology,First Computers,Pentafour Communications,Jetking,IIS Infotech,Boston Education,SQL Star,Datapro and IBM Learning Services.growing presence of these institutes in cities across India is making it increasingly possible for software developers to obtain certificates and diplomas.

It is also noteworthy that this is a private sector response to a market opportunity,namely the demand for training in specialized skills.In the US and elsewhere,for-profitfirms compete with public institutions,such as state and community colleges.In India,the rapid development of private training institutes testifies to the changed economic climate and the channelling of entrepreneurship into economically productive areas away from mere rent-seeking.This change is also at least indirectly related to the rapid growth of the software industry.

To sum up,we believe the evidence indicates that until recently the excess reserve of trained engineering talent was a significant source of competitive advantage for Indianfirms.On the other hand,it provided limited incentives for firms to economize on the use of skilled engineering talent.Instead,the growth of the industry and high salaries have attracted not only new graduates but also engineers,managers and other professionals from other industrial sectors.Despite this,the explosive growth of the software industry has led to the recognition that there is a shortage of skilled engineers,software professionals and good managers. Moreover,faced with a number of attractive options,including job opportunities overseas,these trained individuals are looking for more money and a more professional and rewarding work environment.This,in turn,entails a variety of organizational changes which we discuss below.

5.The contribution of software to economy-wide productivity growth: organizational improvements and the‘demonstration effect’on otherfirms The advancement of software in India has promoted some related service industries,the most notable of which is the rapid growth of the informal IT sector studied by Kumar(2000a,b).Services such as the provision of IT maintenance, data entry and customization services for domestic users,all virtually non-existent a decade ago,are gaining in prominence.Kumar(2000a)estimates that software and these ancillary industries probably accounted for about0.5million jobs in 1998–1999,afigure that is likely to have increased substantially since.According to a recent survey by NASSCOM,IT-enabled services have shown the strongest upsurge(66%)among all industry segments in India over the previous year and are expected to gross a revenue of US$880million in2000–2001,versus US$530 million in the previous year.

The basic business model underlying the development of these services is similar to that of software,namely largefirms outsourcing a part of their operations to take advantage of the availability of a large English speaking population at relatively low wages.Poor infrastructure in telecommunications constitutes the same impediment to the expansion of these services as it does to the growth of software services.This year the two most promising segments areTable8

Current and projected demand for software developers in IT-enabled services in India

IT-enabled services19982008(projections)

No.Rs,millions Can be Rs,millions

employed employed

Back office operations/revenue97004200260,000190,000 accounting/data-entry conversion

Remote maintenance and support1600650180,000135,000 Medical transcription/insurance claim38001400160,000110,000 processing

Call centres1400400100,00060,000 Database services1000450100,00065,000 Content development55002700300,000250,000 Total23,00098001,100,000810,000 Source:NASSCOM,‘The Software Industry in India:A Strategic Review’(1999).

customer interaction services,including call centres,and content development and animation.Table8shows the projections from a NASSCOM survey which

16

forecasts opportunities for IT-enabled services.

On the whole,however,software creates relatively few linkages with other sectors of the economy.But the industry is fast emerging as an exemplar of both good organizational practices and technocratic entrepreneurship.In this section we examine why this is so,and look at the mechanisms through which similar progressive practices could spread to other sectors of the economy.

While a software professional in India earns only a fraction of the salary of his international counterpart,it is still about20times the national average.This difference in international wages makes softwarefirms in India vulnerable to the loss of middle-level employees to the west,and puts continuous pressure on profitability.

Vulnerability to employee attrition and rising wages have created different problems for the software sector.Increasing wages have caused manyfirms to actively introduce stock options to supplement salaries.Until recently,firms used the overvalued technology stocks to supplement wages,or as incentives to secure loyalty by offering employees stock option plans(ESOP).Privatefirms,such as TCS,competed similarly for talented people by offering bonuses calculated on a notional stock price of the company based on its economic value added.Kumar (2000b)estimates that as many as41Indian softwarefirms offered ESOPs in addition to wages in1998–1999.More recently,Infosys and Silverline have 16Medical transcriptions and other forms of data entry,for instance,simply rely on high school or college graduates who know typing and language skills(for which they are trained).For instance,to transcribe the audio-taped diagnoses of physicians located in the US,Indianfirms rely on software developers trained to understand English spoken with an American accent and who have the ability to type in English.started to offer ADR-linked ESOP to their employees as part of a larger retention strategy.

While rising wages directly affect profitability,attrition creates a different kind of threat—the loss of employee-specific knowledge,the possibility that this know-how is transferred to a competitor and,at times,the loss of credibility with the client.Interviews with US-based clients of Indian softwarefirms also indicated that many considered employee attrition a serious problem(Arora et al.,2000a). Several clients recalled incidents of delay caused by an entire project team leaving in mid-term in response to a more lucrative offer.In addition,when attrition is at the senior level,it is possible that thefirm will also lose some of its customers to 17

the competitor.

However,both of these problems have prompted softwarefirms to explicitly introduce human capital management strategies that are comparable with their international rivals.These include the following.

•An increasing role by softwarefirms for private investment in software education and training for their employees.Leading companies like TCS and Infosys spend between4and6%of their revenues on training.It is estimated that the industry as a whole spent7609man-days per organization or5.5 man-days of average training per person per organization on training(Kumar, 2000b).

•Organizational practices designed to retain the interest and loyalty of em-ployees;attention given to charting a management career path for technical personnel,perhaps thefirst signs of a technocracy.

•Organizational innovations designed to diffuse and combine employee-specific knowledge through different approaches,most notably throughflat hierarchies, team efforts and the use of embodied forms of software knowledge such as tools or,where possible,through the use of licenses and IPRs.

The adoption of ESOPs and the charting of career paths in management for technically-qualified professionals have had at least one unintended outcome. Technically-qualified personnel view entrepreneurship itself in a new light.It has become the logical extension of managerial tasks they would have handled in any case as their careers advanced,albeit now with an added element of risk that nevertheless promises large rewards.Thus,ex-employees of HCL and Wipro have become active entrepreneurs,and TechSpan,NIIT,Pertech Computers,Global Infotech,InfoTech Enterprises,STG and Infogain were all set up by ex-Wipro

17An early example of this was the setting up of Infosys by a group of employees who left Patni Computer systems.In recent years,as we document later in this section,this type of spin-out activity has become much more frequent.employees.Similarly,product-based ventures such as Jamcracker,Microland,e4e Ventures,Tarang Software,iLantus,Jumpstart,Qsupport and Mindtree Consulting have been set up by ex-Wipro employees.Other leading softwarefirms are aware of these concerns.Satyam has sought to avoid attrition by investing in corporate venturing.Recent plans announced by TCS indicate that they,too,will start a venture capital fund to encourage start-ups by employees with innovative ideas. In software and other service sectors,shifting into entrepreneurship is also relatively easy because of the low capital requirements and the large venture capital interest in India.Apart from individual incentives to technological personnel for greater private rewards,the social value of this changed attitude towards entrepreneurship is enormous.This is because technically trained person-nel are,through entrepreneurship,directly involved in attempts to commercialize technology and this in turn results in a transfer of knowledge from universities to more applied realms in industrial manufacture.

Some of this change is already visible in the mushrooming growth of the IT-related services sector,as has been demonstrated by Kumar(2000a).This sector shows evidence of the rewards of education in entrepreneurship and the surprising step-up into entrepreneurship amongst all sections of the population. With the advance of private training and the rush among softwarefirms to distinguish themselves in terms of domain expertise,it is becoming relatively easy for professionals from other service sectors to migrate into IT and software through some small investments in software training.Indeed,as the salaries of technologically qualified personnel rise in India,this is one area of substitution employers will be actively looking toward.As softwarefirms seek to build domain expertise and move up the value chain,there is an increase in the demand for marketing personnel,preferably with experience of the US markets.Salaries for software marketing executives rose25%in the last year(Hindu Business Line, 2001),which in turn creates pressures for other service industries(banking and finance,hotels,retail trading)to offer similar rewards,be it ESOPs,or employee friendly policies.Thus,last year the hiring of managers(as opposed to technically qualified persons)in the software sector grew by an incredible30%.At the same time,firms in the banking and insurance sectors(including,in some cases,public sectorfirms)announced decisions to offer ESOP packages to their employees. Even without the luring away of personnel to the software sector,the service sector is one to which the software model can,in theory,be easily transposed.The generic model underlying all service industries is the employment of personnel whose capacity to render a service can be fully utilized.Physical capital requirements of these industries are minimal.Good employees induce customer loyalty and,if undervalued,can leave the company to set out on their own.A valued employee can exert the same pressure on his employer as a capable software programmer.The software model can be almost identically replicated for services that are to be delivered globally.Thus,call centres,data entry and data transcription services operate a generic business model very similar to softwareservicefirms.Given the vastness of the Indian domestic market,in national markets similar elements also apply.

One might ask with some scepticism what are the possibilities for similar qualitative changes to evolve in the older manufacturing sectors of the economy (e.g.steel automobiles,textiles,cement,etc.),where major companies still dominate production and employment,but without the same upward pressure on wages.The surprising evidence is that older businesses are rushing to accommo-date some of these changes in organizational style and employee participation. Mechanisms for the diffusion of organizational practices are,however,different from those in the services and IT-related service sectors.An increasingly large number of the big business houses(Dalmias,Birlas,Mahindras,ITC,Wipro) invested in IT divisions in the early1990s,given the high profitability and relatively low capital needs of software services business at that time.Often these divisions enjoyed considerable autonomy to develop their own management styles, rewards and compensation to employees.The methods and practices that evolved kept pace with competition,but often also deviated from established norms of the parent companies.The struggle of the new and better practices being adopted is,in

18

itself,an interesting story,which is beyond the scope of this paper.However, once the new norms were established in one division of the company,it was rather difficult,for reasons of parity,to inure other divisions from following the same. External borrowing has also played a big role in this process.One of the surprising outcomes of the liberalization of capital inflows in India resulted not only in large amounts of foreign investment,as policymakers had intended,but also in a huge accumulation of foreign debt due to corporate borrowing.As Indian firms tapped the foreign equity markets for theirfinance requirements,equity market norms were being closely followed,particularly with regard to good corporate governance.Thus,despite their relatively low investment requirements,

19

softwarefirms listed on US stock exchanges and pioneered this fund-raising route for Indianfirms.This implied the use of the American system offinancial reporting(GAAP),quarterly reports and close attention to the investor community and their representatives,the equity analysts.Indianfirms were anxious to comply, as these practices signalled their competence and enhanced their visibility to potential customers.Over the next2years,another15Indianfirms will be added to the New York Stock Exchange.Reports indicate that most of these are not from the software sector,but rather from investment-heavy sectors such as pharma-

18As examples we can mention recent events such as the struggle at DSQ started by external investors for professional management to have effective control;the recent attrition among executive staff at Wipro in response to a slow dilution of thefirm’s equity shareholding to employees;the efforts to overcome the stalemate in Mahindra-BT in1995–1996when the company threatened to close by inducting professional management that restored profitability and credibility.

19Infosys(Nasdaq),Satyam Infoway(Nasdaq),Silverline Technologies(NYSE)and Wipro.

A.Arora,S.Athreye/Information Economics and Policy14(2002)253–273271 ceuticals and telecommunications,that will follow the same norms and trans-parency of procedure.

In sum,the diffusion of good management practices from the software sector to other economic areas has come through two channels.As human capital-intensive service sectors provide the software industry with much needed middle-level management,the pressures apparent in the software sector will,in time,ensure that better management practices are adopted in other sectors as well.Although pressures on the manufacturing sector to adopt good management practices do not come from the rise in salaries,nevertheless the participation of old economy capital in the newly emergent software sector has ensured that old businesses are not inured against the adoption of such practices.This trend has been reinforced by thefinancial liberalization of the economy and its opening up to externalfinance.

6.Conclusions and some policy implications

In this paper,we have argued that the rapid advance of software exports from India is a consequence of the country’s comparative advantage in the production of these services rather than the result of an absolute advantage in terms of wage differentials alone.This comparative advantage,in turn,is based on the availabili-ty of a large English-speaking and technically qualified workforce as well as the relative disadvantage in manufacturing due to poor infrastructure investments in the past.Indeed,the larger the size of this disadvantage,the greater the benefit of specializing in software.The transfer of resources from poor productivity sectors to those with greater productivity will also spur the growth process.The last decade has seen higher rates of income growth in the south and west of the country,the areas of software concentration.

These factors favouring India’s software exports are further reinforced by the low domestic demand for similar types of products and services.Though there are gains from specialization,the distribution of gains from such specialization, however,remains uneven due to the structure of the Indian software industry that still remains largely undifferentiated.The latter is,however,a circumstance that will change as the industry matures.

Due to the undifferentiated and service nature of Indian softwarefirms, however,human capital has grown in importance to levels that were hitherto reserved for the scarcerfinancial and physical capital in Indian industry.In an extremely competitive international software market,Indianfirms emphasize the quality of their procedures and human resources to gain a competitive advantage. This microeconomic response has had both positive and negative consequences. One benefit stems from the appreciation of training in an environment where qualified software programmers are scarce and are becoming scarcer.This,in turn, has resulted in substantial private investment in the provision and acquisition of tertiary training.The adverse consequence is the slow hiking-up of prices for the272A.Arora,S.Athreye/Information Economics and Policy14(2002)253–273

most technically-qualified workers,viz.engineers.Engineers had for years supplemented their engineering degrees with other qualifications in order to be accommodated in the better paying management cadre.Currently they are in the enviable position of being able to command large salaries for tasks for which their training is rather irrelevant.

The government has responded to this crisis in engineering education by targeting resources to the engineering colleges and institutes for information technology—the same areas favoured by private investment.Our analysis suggests that the economy as a whole can realize greater benefits by complementing private investments with public investments in basic and secondary education and in physical infrastructure,factors that had constrained the growth of the software industry in thefirst place.The current excess demand for engineers will,in time, be corrected by a more rational and economical strategy,viz.training science graduates in software programming.

The hiking-up of wages for Indian engineers and technically trained labour,and the increased demand for them in the international market have also forced Indian softwarefirms to adopt organizational practices comparable to those of their strongest competitors in the US and elsewhere.This has included the charting of career paths for technical personnel and systems of rewards that share both profits and risks.Increasingly,firms are being forced to provide more congenial and satisfying work environments.We have argued that some of these measures are spilling over into other service sectors,where human capital is more important than physical capital and from where the software sector may draw employees to man middle-level management positions.In turn,this has forced various com-panies,which have controlled large parts of the economy,to adopt more professional and transparent practices.These changes are likely to be slow,but there are other institutional forces at work,most notably the increasing importance of equity markets,and increased competition from imports and from multination-als.

However,greater specialization in software production and exports in the domestic economy(with or without differentiation of the industry)are constrained by precisely the same factors that accounted for India’s relative disadvantage in manufacturing,namely poor and inadequate investment in physical infrastructure, communications and basic education.Thus,while software development offered a window of opportunity because of the large supply of underemployed engineers,it is a narrow window.Furthermore,software has not become the leading sector in India’s economic growth,at least partly because of its poor linkages with the rest of the economy.This poor linkage is itself a consequence of the‘service’rather than‘product’nature of the industry,and its external rather than inward orientation,facts lamented by previous analyses of the industry.

In this paper we have tried to show it is precisely these two facts—the (technical)labour-intensive nature of the industry and an environment of global excess demand for such labour-which have altered the attitudes towards humanA.Arora,S.Athreye/Information Economics and Policy14(2002)253–273273 capital formation at thefirm level and in the national economy.In thefinal analysis,it is in terms of this slow change in attitude towards education, entrepreneurship,and the value of human capital in the economy where the software industry has made its greatest contribution to India’s economic develop-ment.

Acknowledgements

Research for this paper was undertaken when Dr.Athreye taught in the Department of Humanities and Social Sciences at the Indian Institute of Technolo-gy,Delhi,and she gratefully acknowledges the institutional help provided during this time.

References

Arora,A.,Asundi,J.,Fernandes,R.,2000a.Supply and demand for software developers in India.

Heinz School of Public Policy and Management,Carnegie Mellon University,Pittsburgh,Mimeo. Arora,A.,Arunachalam,V.S.,Asundi,J.,Fernandes,R.,2000b.The globalization of software:The case of the Indian software industry.Report submitted to the Sloan Foundation,June2000.

Arora,A.,Arunachalam,V.S.,Asundi,J.,Fernandes,R.,2001.The Indian software services industry.

Research Policy,in press.

D’Costa,A.P.,1998.Technology leapfrogging:software industry in India.Paper presented at the2nd International Conference on Technology Policy and Innovation,Calouste Gulbenkian Foundation, Lisbon,August3–5.

Heeks,R.,1996.India’s Software Industry:State Policy,Liberalization and Industrial Development.

Sage Publications,New Delhi.

Hindu Business Line,2001.A Hard Drive on the Software Front,1January.

IMF,various years.International Financial Statistics.IMF,Washington,DC.

Kumar,N.,2000a.New technology based small service enterprises and employment:the case of software and related services industry in India.International Centre for Development Research and Cooperation,New Delhi,Mimeo,version2.3.

Kumar,N.,2000b.Developing countries in the international division of labour in software and service industry:lessons from Indian experience.Research and Information System for Developing Countries,New Delhi,Mimeo,version2.1.

Lakha,S.,1994.The new international division of labour and the Indian software industry.Modern Asian Studies28,381–408.

NASSCOM,1999.The software industry in India:a strategic review.National Association of Software and Service Companies.New Delhi,India.

Ramarao,P.,1998.Reshaping postgraduate education and research in engineering and technology.In: Review Committee of the AICTE on PG Education in Research and Development in Engineering Technology of the Government of India,pp.126–127.

Raymond,E.,1999.Cathedral and the Bazaar:Musings on Linux and Open Source by an Accidental Revolutionary.O’Reilly Press,California.

Santana,R.,2001.Can Russia’s techies mimic India.Newsweek,20February.

Sengupta,S.,2001.TCS,Infosys,Wipro to outsource from China.Economic Times,5February.下载本文