Out Sour Sing India

8/6/2019 Out Sour Sing India

1/35

ECONOMIC DEVELOPMENT STRATEGIES AND MACRO-

AND MICRO-LEVEL HUMAN RESOURCE POLICIES:

THE CASE OF INDIAS OUTSOURCING INDUSTRY

SAROSH KURUVILLA and ARUNA RANGANATHAN*

This detailed case study o Indias outsourcing industry illustrates the challengesin linking macro and micro human resource policies with an economic developmentstrategy based on export-oriented services. The rapid expansion in the outsourcingo services to India has raised the possibility that this sector will be a key engine oIndias economic growth. Based on extensive feld research carried out over a our-year

period, the authors o this study argue that our interrelated human resource policychallenges threaten the outsourcing industrys growth: two macro problems (cur-rent skill shortages and the inability o the country to produce higher levels o skillsor the long-term growth and sustainability o the industry), and two micro problems(very high levels o employee turnover and rapidly increasing employee costs). Theauthors evaluate current policy responses and suggest options.

*Sarosh Kuruvilla is Proessor o Comparative Indus-trial Relations and Asian Studies at Cornell University,and Aruna Ranganathan is a doctoral candidate at theSloan School o Management, Massachusetts Instituteo Technology. This research is based on feldwork inIndia done by the frst author in 20042005 as a Ful-bright research ellow, with ollow-up visits in 2005 and2006, and feldwork done by the second author in thesummer o 2007. The authors thank Clete Daniel oran introduction to Henry Fords ideas, Mingwei Liu orresearch assistance, seminar participants at the IndianInstitute o Management (Calcutta), the Institute or

Work and Employment Research at MIT Sloan Schoolo Management, the Department o Collective Bargain-ing at Cornell University or comments and suggestions,and our myriad inormants (industry proessionals,academics, journalists, and government ofcials) in theoutsourcing industry or their perspectives.

More details about interview transcripts and codingcan be obtained by contacting the frst author at [email protected].

T he growth in the outsourcing o a variety o low- and high-skill servicesector jobs rom the United States to Indiahas stimulated debate in the United States.Critics o such outsourcing argue that it mayharm the American middle class and seriouslyundermine the economic uture o the coun-try (see Blinder 2007; Greider 2007; Gomoryand Baumol 2001). Proponents argue thatsuch outsourcing will increase profts, expandexports, raise dividends, and result in greatereconomic efciency (or example, Greene

2005). No such ambiguity is present whenthis phenomenon is viewed rom an Indian

perspective, where the outsourcing industryhas come to be viewed as the primary engineo the countrys development over the nextew decades, contributing broadly to GDPgrowth, employment growth, and povertyalleviation (Srinivasan 2006:204). Indiasreputation as the business process outsourc-ing capital o the world raises the question owhether an economic development strategybased on export-oriented services (EOS) isa viable option or developing countries, in

contrast to the low-cost, labor-intensive ex-port-oriented industrialization(EOI) strategy


2/35

40 INDUSTRIAL AND LABOR RELATIONS REVIEW

that has been the hallmark o Asian develop-ment (World Bank 1993).

Many have highlighted the linkage

between successul EOI strategies and ap-propriate labor policies (see Amsden 1989;Deyo 1987, 1989; Kuruvilla 1996a). In thispaper, we ocus on the linkage between anexport-oriented services strategy (EOS) andnational human resource policies (particu-larly skills development policies). Specically,the policy question we investigate is whetherexisting work orce skills and national hu-man resource policies in India will supportor retard the development o this sector,

given that the quality o human resourcesis a critical actor in determining compara-tive advantage (or example, Wood 1994;Godrey 1997). Thus, although we ocus onIndia, this issue has relevance or the manydeveloping countries (or example, Hungary,the Philippines, Malaysia, the Czech Repub-lic, China, and Brazil) that seek to emulateIndias success in developing service sectoroutsourcing strategies.

Although the availability o the worlds

largest pool o low-cost, English-speak-ing, scientiically trained manpower hasbeen Indias major source o competitiveadvantage and success in outsourcing, theindustry now aces critical labor shortages.Further, we argue that both near-term andlong-term growth prospects o the Indianoutsourcing industry hinge crucially on itsability to overcome a series o interrelatedmacro (national-level) and micro (irm-level) human resource policy challenges.

The macro challenges include current skillsshortages and a deicit in the productiono high-end skills essential to the long-termsustainability o the industry, and the microHR policy problems are the extremely highlabor turnover and rapidly mounting humanresource costs. We use the skills equilibriaramework (Finegold and Soskice 1988)to evaluate current eorts to solve theseproblems and to suggest policy solutions. A key implication o our analysis is that

there is an urgent need to reexamine thecurrent roles and policies ollowed by thegovernment the industry association and

Methodology

Since the purpose o our research was to

identiy the changes in and development oa brand new, growing industry, we relied ona variety o methods to gain understanding.Our primary source o perspective was79 interviews spread over the 20042007period. These interviews, conducted inan unstructured ormat, averaged about 2hours in length, though some took as muchas hal a day and others lasted only an hour.Interviewees included outsourcing industrymanagers (both operational and HR) and

other experts (proessors, journalists, andcritics). We used a snowball sampling method,obtaining interviews not just with a pre-set listo persons, but also with others suggested tous by the interviewees themselves as persons who might have dierent perspectives onthe sector or dierent experiences manag-ing human resources in their rms. Mosto our inormants were rom larger rms(large rms dominate in terms o industryrevenues), and to that extent there is a large

rm bias in our sample.Together with interview data, we use

data rom the rst authors observation oour workplaces in the business process out-sourcing (BPO) segment o the industry in2004. Typically these observation sessionslasted roughly hal a shit. In two cases thework fow was observed, giving us a sense othe complexity o the work involved in thetypical low-end business process rm. Theobservation o the work process conrmed

the conclusion o much prior research (orexample, Batt et al. 2005; Remesh 2004)that low-end BPO and call center work istypically Tayloristic in character. In theother two cases, the observation ocused onthe training o new hires, particularly in theaccent neutralization o call center opera-tions. Observation o these processes wasalso buttressed by short interviews with thetrainers. In addition, in each o these rms, we examined rm-level human resource

inormation (particularly the demographicprole o employees, HR cost structures, HRpolicies and turnover data)


3/35

INDIAS OUTSOURCING INDUSTRY 41

data, we identifed the important problemsacing the sector. When in doubt about theimportance o certain problems, we went

back to selected interviewees in 2005 and2006 (although some o them had changedtheir jobs during that period). Since mucho the studys ocus is on frm-level HR prob-lems and issues, the second author spent twomonths embedded in the human resourcedepartment o one o the three leading out-sourcing frms in 2007. As a participantobserver, the second author was able tostudy personnel policies and practices insome depth. Based on interviews with HR

managers attending the annual HR Summitor the industry in July 2007, we believe thatalthough this irms policies were those oa leader and trendsetter, they have somegeneralizabilityas suggested, or example,by the similarity o measures employed tocombat turnover.

Our understanding o the problems acingthe industry was considerably enhanced byclose perusal o trade magazines. Whilethe most important trade journal was Data-

Quest, a number o internet-based sourcesare also widely used by outsourcing industryemployees. These include daily technologynews websites like Wiredas well as dedicatedinormation technology (IT), telecommuni-cations (Telecom), and BPO news resourceslike InfotechOnline and a number o blogs(or example, TCP BPO is a blog specif-cally catering to BPO workers in India). Inaddition, most leading newspapers in Indianow have an InoTech section dedicated

to industry news.To get a sense o the contours o the

outsourcing industry, we relied on datarom NASSCOM (the National Associa-tion o Sotware and Services Companies).NASSCOM unctions as the trade body andchamber o commerce o the industry, andis the onlycomprehensive source o industrydata, which are collected via membershipsurveys and through special studies done byNASSCOM in partnership with consulting

frms such as McKinsey (or example, theNASSCOM-McKinsey reports) and Hewitt(or example the NASSCOM-Hewitt rewards

with data rom CMIE (the Center or Moni-toring the Indian Economy).

Although one should be appropriately

skeptical o industry data provided by anyindustry association, NASSCOMs researchwing has established a reputation or pub-lishing authentic and credible data (Joshi2002). Because o the level o detail providedby NASSCOMs data, which allows or analy-ses o trends over time, most academiciansand observers preer NASSCOM over othersources, such as the Department o Electron-ics, the Reserve Bank o India, and the CMIE.Tharakan, van Beveren, and Van Ourti (2005)

cross-checked the accuracy o NASSCOM datawith data rom 51 national IT associations,45 national statistical ofces, and 9 interna-tional data sources and concluded that theNASSCOM statistics were comparable, butmore comprehensive. Most managers inter-viewed by Arora and Asundi (1999) regardedNASSCOM fgures as accurate and credible.Adding to their credibility is the act, as notedby Joseph and Harilal (2001:16), that theyare compiled rom frm-level data and that

there is hardly any sotware or BPO companyin India which employs more than 20 proes-sionals that is not a member o NASSCOM.Hence in this paper, the various fgures onindustry size and growth presented were as-sembled by us rom a variety o NASSCOMreports and studies.

Companies in the industry belong toNASSCOM or a variety o reasons, mostlyhaving to do with the ability to network (bothdomestically and with the international sot-

ware community), to lobby the government,and to gain access to the variety o servicesthat NASSCOM oers its members. Theseservices include research on sotware tech-nologies and market intelligence services(including new emerging opportunity areas),training in industry best practices, knowledgeabout quality and security standards/mini-mum requirements, access to the NationalSkills Registry, and subscription to weeklyand biweekly publications (like HR Connect,

Communique, and BPO Newsline) as well asannual reports (like the Strategic Review).NASSCOM also organizes seminars human


4/35


sessions, training and mentoring workshops,leadership orums, and other events, basedon industry trends and member needs. In ad-

dition, NASSCOM plays the role o strategicadvisor to member companies by mentoringthem on world-class management practices.Thus, there is a strong incentive or most rmsin the sector to become members o NASS-COM, and its 1,200 members constitute 95%o the revenues o the industry and includeboth Indian and multinational rms. Hence,it is a representative organization.

NASSCOM is also infuential (a relevantissue in this paper, especially when, in the last

section, we discuss policy options). Althoughits initial charter was to lobby and obtainconcessions rom the government on taxmatters, NASSCOMs biggest achievementis that it has at least managed (through avariety o methods) to put various issues on aslow-moving governments policy agenda. Ithas successully lobbied or industry-riendlypolicies such as zero tari protection, strongintellectual property and data protectionlaws, stepped-up deregulation o the telecom

market, the creation o sotware technologyparks, and private sector participation in theeducation system (all o which have helpedthe growth o the industry). Moreover, unlikeother chambers o commerce, NASSCOM hasa voice in the creation o government policy,since it now has representatives in variouscommittees within government departments,including the Ministry o Inormation Tech-nology, Ministry o Commerce, Ministry oFinance, Department o Telecommunication,

Ministry o Human Resources Development,Ministry o Labor, and Ministry o ExternalAairs. A nal indication o its infuence is theconsulting services it provides to state govern-ments to advise them with regard to telecominrastructure, physical inrastructure, and ITorientation. Thus, NASSCOM plays a centralrole in representing the industrys interests inthe political arena, providing a portolio orelevant services to its members and consult-ing services to state governments.

The Indian Outsourcing Industry

tion TechnologyInormation TechnologyEnabled Services). The IT-ITES industry issubdivided into our segments: IT Services

and Sotware (henceorth called the sotwaresegment), ITESBusiness Process Outsourc-ing (henceorth called the business process-ing segment or BPO), Hardware (which isrelatively small, and domestically ocused,and thereore ignored in this paper) andEngineering Services, R&D, and SotwareProducts (a newly created segment hence-orth called R&D). Note, however, that thelines between some o these segments areincreasingly becoming blurred as a result o

the changing nature o outsourcing workas well as industry consolidation.

Origins of the Software andBusiness Process Outsourcing Segments

Prior research has identied a number ocatalysts that have infuenced the growth othe sotware segment. (There is much debateabout the relative importance o each o thesecatalysts; or diering views, see Srinivasan

2006; Arora and Athreya 2002; Saxenian 2002;Dossani 2006; Dossani and Kenney 2003; andNilakant 2006.) The catalysts include thegeneral liberalization o the Indian economy;government policies avorable to sotware andtelecommunications; entrepreneurial Indianrms (such as TCS, Wipro, and Inosys) thatexploited the opportunity in the 1980s and1990s to export highly trained engineers tothe United States to do programming workor U.S. rms (a practice known as body

shopping); the Y2K crisis, which popularizedIndias reputation in sotware; the Indiandiaspora in Silicon Valley; and, nally, Indiaslarge pool o low-cost, skilled, English-speak-ing sotware proessionals. The reputationdeveloped by Indias sotware segment bythe 1990s, in turn, was a key catalyst in thedevelopment o the BPO segment in the late1990s and early 2000s, along with the avail-ability o telecommunications bandwidth atlower prices, technological developments

such as the ability to digitize documents,a growing opposition to body shoppingpractices in the United States and global


5/35


processing work. The pioneering eorts olarge MNCs such as GE, American Express,and British Airways in setting up outsourcing

operations in India to capitalize on the hugesalary cost dierentials between similarlytrained employees in India and the UnitedStates paved the way or more frms to emulatetheir practices (Dossani 2006).

Industry Size and Growth

The outsourcing sector is currentlyvalued at $47.8 billion (20062007) and isprojected to grow to $150 billion by 2012.

Table 1, which provides both value and growthdata, shows that while the whole industryhas grown steadily, the growth rates o thedierent segments vary. Although the sotware segment grew by 33% in 20062007,the annual average growth rate during the20002007 period was only about 28%, com-pared to the BPO segments explosive annualaverage growth rate o 53% during the sameperiod. And while the sotware segment hasaccounted or the lions share o total industry

revenue50% in 20062007, compared with20% or the BPO segmentprojections basedon current (2007) growth rates show the BPOsegment overtaking the sotware segment by2012. The relatively new R&D segment grewat 3035% per year between 2005 and 2007,albeit rom a small base.

Table 2 indicates the signifcance o theoutsourcing industry to Indias economicdevelopment. While the overall contribu-tion o the outsourcing industry to Indias

GDP is still relatively small, accounting oronly 5.4% o GDP in 20062007, projections1suggest a fgure o 12.3% by 2012. This isremarkable considering that the agriculturalsector (Indias largest sector, employing about60% o its population) currently (2007) ac-counts or only 18.5% o the countrys GDP

(Ministry o External Aairs, Government oIndia).2 The signifcance o the industry isbetter revealed by its contribution to export

earnings. It accounted or more than 25% oIndias export earnings in 20062007. (Thesotware segment was responsible or 58% othe industrys export earnings contribution,while the BPO segment accounted or 27%.)The industrys contribution to employmentis modest in a country as large as India, butsignifcant in absolute terms. The numbero people directly employed by the industrystood at about 1.6 million in 20062007 (pro-jected to be 5 million by 2012), and indirect

employment includes about 1.2 million jobsin ancillary services such as transport, cater-ing, construction, security, and housekeeping(NASSCOM 2005).

The dramatic growth rate o the industry,its relatively large contribution to exportearnings, and the potential or signifcantcontributions to GDP as worldwide demandor outsourcing increases (see Blinder 2007or some estimates o the number o U.S.service sector jobs prone to outsourcing)3

have placed this industry in a good positionto become Indias most important develop-ment engine.4

Industry Composition and Structure:The Variety of Work Outsourced

Sotware Services Segment

Thesotware services segment can now bedivided into three dierent service lines.

1These are estimates rom Gartner, a consulting frmthat ollows this industry. The literature is replete withsuch projections. The frst authors analysis o variousprojections made since 1991 suggests that while there

are dierences in the numbers, usually the actual growthrate occurring over the two years ollowing a projectionhas exceeded the predicted rate. We later argue that

2To provide a comparative perspective, the largestcontributor to U.S. GDP is the inormation industry, at7.4% (Bureaus o Economic Analysis, U.S. Departmento Commerce).

3The Americas (where o-shoring is growing at anannual rate o more than 20%) account or 67% o allIndian exports in this industry, while Europe (whereo-shoring is growing at a slower rate, with the excep-tion o Britain) accounts or 25%.

4Clearly this industry is important. But in a largecountry like India, it is oten a mistake to attribute too

much signifcance to one industry. The Bharatiya JanataParty government lost the 2004 elections partly becauseit highlighted the outsourcing industrys growth in its


6/35


Table1.GrowthofIndias

OutsourcingIndustry.

Total

Total

Software

Software

BPO

BPO

Hardware

Hardware

R&D

R&

D

Industry

Industry

Segment

Segment

Segmen

t

Segment

Segment

Segment

Segment

Segm

ent

Value

Growth

Value

Growth

Value

Growth

Value

Growth

Value

Grow

th

(US$billions)

Rate

(US$billions)

Rate

(US$billions)

Rate

(US$billions)

Rate

(US$billions)

Rate

12.1

48%

7.8

47%

1.0

66%

3.4

42%

NAa

NA

a

13.4

11%

8.7

12%

1.6

60%

3.2

6%

NAa

NA

a

16.1

21%

9.9

14%

2.7

68%

3.4

6.2%

NAa

NA

a

21.6

34%

10.4

5.5%

3.4

26%

5.0

47%

2.9

28.4

31%

13.5

30%

5.2

53%

5.9

18%

3.9

34%

37.4

32%

17.8

32%

7.2

38%

7.0

19%

5.3

36%

47.8

28%

23.7

33%

9.5

32%

8.2

17%

6.5

23%

62

24%

28

9%

21

64%

13

29%

NAb

NA

b

148

46%

55

32%

64

68%

29

41%

NAb

NA

b

rce:NASSCOMS

trategicReview(2007).

wsegmentcalledEngineeringServices,R&D,andSoftwareProductscreated(earlierencompassedbysoftwareandBPOsegments).

ojection

s(madebeforethecreationofthe

newsegment).


7/35


Table2.Contribut

ionoftheOutsourcingIndustrytoIndiasGDP,

Exports,and

Employment.

Services

Contrib.

Contrib.

Total

Grow

th

Exports

Growth

Growth

Growth

to

toIndias

Services

Rate

of

as%of

Software

Rateof

BPO

Rateof

R&D

Rateof

empl.

GDP

Exports

Services

TotalExport

Exports

Software

Exports

BPO

Exports

R&D

(millions

(%)

(US$billions)a

Exports

Earnings

(US$billion)

Exports

(US$billion)

Exports

(US$billion)

Exports

ofworkers)

2.7%

6.2

56.9

%

13.9%

5.3

55.6%

0.9

64.6%

NAb

NAb

0.4

2.9%

7.7

24%

17.6%

6.2

16.4%

1.5

60.8%

NAb

NAb

0.5

3.2%

9.6

25%

18.2%

7.1

14.6%

2.5

67.2%

NAb

NAb

0.7

3.5%

12.9

34%

20.2%

7.3

2.8%

3.1

24%

2.5

0.8

4.1%

17.7

37%

21.2%

10.0

37%

4.6

48%

3.1

24%

1.1

4.8%

23.6

33%

23.0%

13.3

33%

6.3

37%

4.0

29%

1.3

5.4%

31.3

33%

25.0%

18.1

36%

8.3

32%

4.9

23%

1.6

rce:NA

SSCOMS

trategicReview(2007).

talServ

ices=Software+BPO

+R&D.

wsegm

entcalledEngineeringServices,R

&D,andSoftwareProductscreate

d(earlierencompassedbysoftwareandBPO

segments).


8/35


(Table 3 reports value, growth, and revenuestatistics or each service line.) Each serviceline comprises both a large and stable low-end marketthe bread and butter o most

frmsand a small but growing high-endmarket. The largest service line, Project-Oriented Services, accounts or 56% o totalsegment revenues, and is dominated by thelow-end customized application and mainte-nance o sotware or client frms, which hasbeen the mainstay since the early 1990s. Olate, there has been growth in higher-end,highervalue-added services such as systemintegration (where growth rates averaged40% in the last two years) and IT and Net-

work consulting (see Table 3), evidence thatthe industry is moving up the value chain.An area o growth o particular importancehere is consulting in enterprise applicationintegration, which is key to designing globalvalue chains and building eective e-businessinrastructures. These highvalue-addedservices (which require good postgraduateskills) have propelled Indian frms such asInosys to establish consulting units in theUnited States to compete with global leaders

such as Accenture and IBM.The second service line, IT Outsourcing,

which accounts or 33% o the segments

services or all clients (not customized orany single client) and encompasses bothlow-end activities such as applicationsoutsourcing (which is the largest activity)

and high-end services such as network inra-structure management. Typical exampleso such work are cases in which Westernclient companies (or example, Bell-South)outsource the entire gamut o IT applica-tions to Indian irms.

The Support and Training service line ismore heavily dominated by the low-endoperations o sotware support, such as sot-ware support to a domestic U.S. consumer.However, there is also a high-end activity

herethe export o IT education and train-ing services. A notable example is EducompDatamatics, which has developed its SmartClassroom or U.S. schools.5

Table 3. Growth by Service Lines in the Sotware Segment, 20032004.(billions o U.S. dollars)

Service Line FY 03 FY 04 Growth

Project-Oriented Services 3.23 3.85 19.2%

IT Consulting 0.08 0.12 50.0%System Integration 0.10 0.14 40.0%Custom Application Development and Maintenance 3.02 3.54 17.2%Network Consulting and Integration 0.03 0.05 66.7%

IT Outsourcing 1.94 2.45 26.6%

IS Outsourcing 0.01 0.02 100.0%Application Outsourcing 1.85 2.16 16.8%Network Inrastructure Management 0.08 0.27 260.0%

Support and Training 0.37 0.61 64.9%

IT Training and Education 0.02 Hardware Support and Installation 0.02 0.04 100.0%Packaged Sotware Support and Installation 0.35 0.55 57.1%

Total 5.54 6.91 24.8%

Source: NASSCOM, IDC (2006).

5A Smart Classroom is already up and running atFranklin Public School in Santa Barbara, Caliornia. A knowledge center set up at the school acts as thenerve center o the network. It is connected througha campus-wide network to the existing classrooms, which simply turn into ETECs (Educomp Technol-

ogy Enabled Classrooms). The teacher then uses thedigital resources, such as animation, graphics, images,and video clips, while teaching the chosen topic in the


9/35


In sum, the composition o work donein the sotware segment is slowly but surelychanging rom low-end work to increasinglyhigh-end work, with the high-end work evi-dencing aster growth (see the last columno Table 3). This change has brought abouta transormation in the business strategies oIndian rms, which are moving rom simpleo-shore outsourcing o some elements oIT to a more evolved sophisticated global

delivery model. Hence, on the one handwe see more and more IT services being sentoshore to India (the oshore percentage osotware revenues increased rom 43% in 2000to 71% in 2005), and on the other hand In-dian rms are providing higher value-addedservices onsite (that is, in the United States),evidenced by the opening o consultationlocations in the United States. Inosys, orexample, has a 1:3 onshore-oshore ratio, sothat or its 9,000 U.S.-based employees hired

to serve as an interace to the U.S. clients,there are 27,000 oshore employees.

Thus, the Indian sotware segment isbecoming more skill-intensive, consistentwith Krishnan and Dayasindhus (2004) sug-gestion that the industry must develop newlayers o competitive advantage, given theincreasing competition in sotware servicesrom China, Hungary, Argentina, Latvia, Ma-laysia, Estonia, Slovenia, and Romania (Nairn2004). However, this movement requires

that the country produce the skills required.As we will demonstrate in the next section,the Indian university system has little or no

the skills o system architects, researchers, andvisionaries), and even skills lower down thechain, such as the global domain knowledgeneeded by business managers, are already inshort supply (interview, Narayan).

The Business ProcessOutsourcing (BPO) Segment

The reasons or Indias dominant position

in the global market or business processoutsourcinga 4850% market share in2004 (NASSCOM 2005)are refected in theadvice and recommendations corporationsturn to when orming their oshore locationstrategies. The most heavily used resourceo this kind is a comprehensive evaluationtool developed by the consulting rm ATKearney. AT Kearneys analysis indicates thatIndia has a clear lead over its major competi-tors, particularly with regard to cost and the

quantity and quality o skills. Indias scoreon the index is a 7.12, compared to scoresranging rom 5.33 to 5.61 or the other BPOcompetitors (China, Malaysia, the Czech Re-public, Singapore, the Philippines, Canada,Chile, and Poland).6 Indias cost advantagevis--vis the United States is demonstrated bya comparison o call center costs between thetwo countries (Table 4).

Table 4. Comparing Cost Structures in the U.S. and Indian Call Center Segments.

United States India Saving

Cost Element Valuea

Share Valuea

Share On U.S. Cost BasePersonnel Costs 42,927 73.3% 6,348 48.4% 85%IT/Telecom Costs 2,400 4.1% 3,770 28.7% 57%Oce Facility Costs 3,700 6.3% 1,991 15.2% 46%Other G&A Expenses 9,571 16.3% 1,012.5 7.7% 89%

Total Cost 58,598 100% 13,121 100% 78%aUSD/FTE/Year.Source: NASSCOM, Evalueserve Analysis, Strategic Review (2006).Notes: Blended rate or voice and non-voice operations. The above gures are indicative only, and the actual

cost savings could vary by process. FTE: Full Time Employee. G&A: General and Administration.

6They use an index based on 39 criteria, which are

organized under three broad categories with varyingweights: nancial structure (40%), people skills andavailability (30%), and business environment (30%).


10/35


The BPO segment, growing at an annual

average rate o about 50% since 1999, sawemployment rise rom 50,000 in 19992000 to415,000 by mid-2006. Rather like the sotwaresegment, the BPO segment is characterizedby a large low-end section and a discern-ible and growing high-end outsourcingmarket that involves very advanced skills (owhich many are in short supply). However,there is greater variability in service linesin the BPO segment than in sotware. Themushrooming o newer service lines on al-

most a weekly basis makes it difcult or usto categorize them. Table 5 lists descriptionso the basic service lines or which we havedata or 2005; below, we describe the manynew categories that are emerging.

Customer care (call centers) and supportservices or a number o dierent industriescomprise the largest service line, accountingor close to 40% o the industrys employeebase and a third o its revenues. Typicalexamples include the outsourcing o Delta

Air Lines worldwide reservation center tothe Indian frm Wipro-Spectramind (nowWipro BPO) and Amazons outsourcing o itscustomer service to Daksh (a frm now ownedby IBM). The fnance and accounting serviceline includes the standard business processeso transaction management, general account-ing, treasury and risk management back ofcework, statutory reporting, and compliancework, while the payment services line includesa variety o billing services, insurance claims

processing, and tax processing. The HR out-sourcing service line is experiencing rapidgrowth (this typically comprises payroll and

employee communication services). In ad-

dition, there is a growing medical and legaltranscription service line.

New service lines have emerged in feldssuch as fnance (analysis o economic data,business intelligence management o oreigncommodity accounts, and lower-end servicessuch as home loan processing, mortgageservices, and debt collection and recovery),medicine (creating medical education data-bases, doctors desk reerence guides, andnotes on drug efcacy; more recently, read-

ing and analyzing patient X-rays), printing(pre-publication o scientifc journals), andremote education (such as providing mathtutorials to U.S. schoolchildren), leading tonew service line categories known as FinancialProcess Outsourcing (FPO) and KnowledgeProcess Outsourcing (KPO).

Thus, like the sotware segment, the BPOsegment is moving up the value chain,but more rapidly. At the low end o BPO,revenues are tightly and linearly linked to

manpower employed.7 However, at the FPOand KPO level, it is high skills, not numberso employees, that matter. The dierencein complexity between regular BPO jobsand the jobs in the KPO amily can be seenin the dierences in billing rates o BPOirms. Whereas billing rates or traditionalHR outsourcing average about $15 perhour, the average or a wide range o KPOactivities is about $45 per hour and rising,as shown in Table 6.

Table 5. Growth by Service Line in BPO Segments, 20042005.

Export Revenue Export RevenueService Line (Low End) (US$ millions), 2004 (US$ millions), 2005 Growth Rate

Customer Care 1,200 1,500 25%Payment Services 430 620 44%Finance 835 1300 55% Administration 540 840 Human Resources 75 165 120%Content Development 550 670 22%

Source: NASSCOM Strategic Review (2006).Note: This table contains data rom the traditional low-end BPO operations.


11/35


Engineering Services, R&D, andSoftware Products (the R&D Segment)

This newly created segment (originally parto BPO) represents very high-end servicesthat have the greatest potential or growth,according to industry analysts. Here too thevariety is great and does not neatly t intoNASSCOMs categories. But some o thesejobs require doctoral-level skills or advancedmasters-level skills. Table 7 shows a limitedsubset o the varieties o work done in thissegment, and we discuss selected examplesin more detail below.

In engineering services, a particularlyimportant line, India had 15% o the world-wide market share in 2006, according to astudy by Booz, Allen, and Hamilton (seeNASSCOM 2006). Worldwide outsourcingo engineering services is currently worth$15 billion. The study suggests that Indiaaccounts or 28% o the worlds engineer-ing services outsourcing talent in low-costcountries (Russia and China are the closestcompetitors with 11% and 10%, respectively).

Engineering services span several industries(verticals in industry parlance), includingaerospace, automotive, telecomm, utilities,industrial construction, pharmaceuticals,medical engineering, consumer electron-ics, and industrial chemicals. Aviation andaerospace accounts or a particularly largeportion o engineering services outsourc-ing activities and is growing rapidly. Forexample, Wipro employs 450 engineers inits aerospace engineering services outsourc-

ing and plans to double that in the nextew years. Sathyam Computers has createda new subsidiary, Sathyam Aerospace, toocus on providing engineering services tothe global commercial aviation aerospaceand deense markets. Further, many en-gineering MNCs (or example, Siemens,Snecma Aerospace, Samsung, Philips, Del-phi, Daimler, and Bosch) have establishedengineering research and developmentcenters in India.

In sotware, there is growth in the R&D-ori-ented eld o embedded sotware and systems(which grew by 44% in 20052006) Large

some o their embedded system requirements

rom India, and a growing number o themestablished sotware design centers in Indiain 20052006. As o 2007, over 900 engineerswere working at Motorolas Global SotwareGroup (GSG) at Bangalore and Hyderabad,creating cutting-edge technological solutionsor wireless products and providing embed-ded solutions and services or 3G phones.There is also growth in the research-intensivearea o sotware product development (thehighest end o the sotware value chain).

Several successul sotware products havebeen designed, developed, and marketedby rms located in India; indeed, a ew haveachieved global success (or example, Tal-isma Corporations E-CRM suite, InosyssFinnacle or the nancial and bankingindustries, Ifexs fexcube or the internetbanking industry, Tally Solutions tallyees6.3 accounting and nancial managementpackage, and Cranes SYSTAT statisticalanalysis package).

Apart rom engineering services andsotware, R&D outsourcing is growing par-ticularly in two industries: lie sciences andnance. In the lie sciences industry, thework pertains primarily to medical research,drug discovery, drug testing, clinical trials,and development o search tools to searchgenomic databases. The U.S. Department oLabor estimates that at least 36,000 jobs inthis industry will move to India over the nextew years. In the nancial services industry,

very high-skill jobs in equity research, datamodeling, and actuarial analytics have alreadymoved rom Wall Street toBangalore! (For

Table 6. Service Lines and BillingRates in the BPO Segment, 2005.

Service Line Billing Rates ($US/hr.)

KPO 3075Content Development 1842Human Resources 1630 Administration Finance 1228Payment Services 1228Customer Care 1222

Source: NASSCOM (2005)


12/35


Changing Industry Structures

The rapid pace o change in the natureo jobs being outsourced to India has trans-

ormed industry structure. On the one hand,there is consolidation withinboth the sotwareand BPO segments. In the sotware segment,although there are over 3,000 dierent ITfrms, the structure is pyramidal, with fvevery large frms (> $1 billion in revenues),about 70 medium-sized frms (> $25 millionin revenues), and a large number o bot-tom eeders.8 The top 20 frms accountedor 48% o revenues in 20032004 (up rom46% in 20022003). In the BPO segment

there were more than 400 frms in Indiain 2006, but 10 large frms dominated theindustry.9 Consolidation is also eroding thedistinction between captives (BPO frmsowned by U.S. client frms such as Dell, EDS,and HP that service only their owners) andthird-party frms (both Indian-owned andoreign-owned frms providing BPO services

to any client, and normally servicing a largenumber o clients simultaneously).

In addition, there is cross-segment consolida-tion, blurring the boundaries between the

segments. The nature o the work beingoutsourced has provided an opportunity orincreased synergy between sotware supportand services and BPOs. All the major sot-ware frms have started or acquired their ownBPO units; or example, Wipro started WiproSpectramind, Inosys started Progeon, TCSbought Phoenix, and IBM recently acquiredone o the largest third-party frms, DAKSH,leading to the emergence o true outsourc-ing multinationals. Inosys, or example,

integrates both sotware and BPO work andis widely diversifed, with operations in en-gineering (engineering design), aerospace(airrame design, seating), automobiles,heavy engineering, banking and fnance(transaction processing systems, tradingsotware), healthcare (creation o a nationalhealth record system), logistics (sotware orsupply chain management, transportation),manuacturing (logistical systems or fsher-men), and several other felds.

In sum, the sotware segment, the BPOsegment, and the R&D segment in the out-sourcing industry are experiencing rapid

Table 7. Export Revenue per Service Line in theR&D Segment (Engineering Services, R&D, and Sotware Products).

(billions o U.S. dollars)

Export Revenue, Projected ExportService Line 2003 Revenue, 2010

Basic Data Search Integration & Management 0.3 5.0

Market Research/Business Intelligence 0.1 0.5

Equity Research/Actuarial Analytics/Data Modeling 0.4 0.5

Engineering Design 0.4 2.0

Animation & Simulation 0.2

Medical Content and Services (X-Rays, etc.) 0.1 0.3

R&D (outside o IT) 0.2 2.3

Biotech and Pharmaceuticals 0.28 3.0

Source: Study by Evaluserve, reported in NASSCOM Strategic Review (2005).Note that these are the most conservative estimates we could fnd. NASSCOMs annual business report on the

growth o each o these segments paints a ar more positive picture o growth, although we have not been able tofnd comparative data on the dierent service lines in high-end KPOs.

8The 10 largest sotware frms in 2006, in descendingorder by size, were Tata Consultancy Services (TCS),

Inosys, Wipro, Satyam, HCL, Patni, I-Flex, Tech Ma-hindra, Perot Systems, and L&T Inotech.

9These are, in descending order by size in 2006,


13/35


These developments have revealed manyskill shortages (giving rise to a number ofrm-level problems), which raises the ques-

tion o whether the country can generatethe higherorder skills required to sustainthe headlong growth and transormation othis industry.

Macro and MicroHuman Resource Challenges

We argue, on the basis o our feld inves-tigations, that our critical and interlinkedhuman resource challenges threaten the

near-term and long-term prospects o this in-dustry. Two problems are macro in nature,that is, requiring government-level policyinterventions. The frst macro HR problemis a shortage o skilled human resources, andthe second, perhaps more critical problemis the difculties Indias advanced educa-tion and training inrastructure will ace asit is tasked with producing the high-skilledmanpower the industry will need, given itscurrent trajectory toward more knowledge-

and research-intensive work. Policymakersacknowledge and are trying to address thefrst problem but have not yet seriously con-ronted the second.

The micro human resource problems,themselves interrelated, are also linked tothe macro ones, but they require mostlyfrm-level solutions. The frst problem ishigh average turnover rates, which in 2007were about 20% in the sotware segment and5060% in the business process outsourcing

segment. These high rates are partly relatedto the macro problem o skills shortages, butare also a unction o poorly developed frm-level human resource strategies. The secondand closely related problem is rapidly risinghuman resource costs in the BPO indus-tryparticularly at the low end, where lowcosts are a source o competitive advantage.Cost increases are a unction o turnover,to be sure, but as we will demonstrate, solu-tions will also require some macro policy

interventions. Although the problems oturnover and skill shortages are increasinglyattracting business press coverage this is the

Shortage in Supply of Skilled Labor

A key source o competitive advantage

or the sotware segment in India has beenthe availability o low-cost skilled sotwareengineers, and or the BPO segment, well-educated, English-speaking manpower.This large HR pipeline has been prized byclients and has contributed to Indian frmsreputation or being able to ramp up quickly(interview, Varadarajan). However, there arecritical labor shortages in both segments.These shortages were apparent during thefrst phase o our feldwork in 2004, but are

more widely recognized now. The head oNASSCOM in 2006 suggested that in the nexttwo years India could lose potential o-shor-ing work worth billions due to critical talentshortages (NASSCOM 2006).

Software segment. About 562,000 sotwareproessionals were employed in 2007 (NASS-COM 2007). This is a steep rise rom the398,000 sotware proessionals employedin 2006 and the 297,000 employed in 2005.Demand or sotware proessionals is grow-

ing steadily. For example, the top threesotware frms hired more than 72,000 inFY 20062007. Inosys alone recruited over25,000 in 2006 (Long 2006). However, esti-mates o uture demand vary widely. In 2006,NASSCOM suggested that annual demandor sotware proessionals would outstripsupply by a margin o 70,000 in 20072008and 235,000 by 20092010. Analysis andprojections by Dataquest (a trade journal)suggested a more dire picture, with a shortall

o 300,000 predicted or FY 20062007 itsel.Our interviews with managers in the indus-try present a compelling picture o currentshortages, but a more nuanced view regardingwhere the shortages are occurring.

At the time o our interviews, the manag-ers reported shortages o programmers withcommon sotware skills (or example, pro-fciency in Java, C++, ERP, SAP, and J2EE), aswell as shortages o application developersand database engineers, but they cited a

more severe shortage o workers with morespecialized skills (or example, data ware-housing experts Citrix-certifed enterprise


14/35


tors, Cisco-certifed network proessionals,Microsot-certifed systems administrators,certifed inormation systems security proes-

sionals, and Red Hatcertifed proessionals).Managers also reported shortages o work-ers with managerial skills (or example,systems administrators, project managers,and coordinators/managers o teams o sot-ware proessionals across various countries). Also, urther up the sotware value chain,consultants who can advise clients abouttheir technology needs were reported to bein short supply. Finally, our intervieweesreported a great need or researchers, archi-

tects, and visionaries o the kind or whomthe job market is truly international and whothereore tend to gravitate abroad. As oneresearcher told us, Tenure at a U.S. univer-sity trumps pay (interview, Venkatratnam).The shortage, in the views o the managerswe talked with, appears to be more criticalat the middle and upper management levelsand in high-skill categories.

I the extent o the labor shortage in thesotware segment is difcult to pin down, so is

the supply. Although the business press andpopular press tout the production o 350,000engineers annually (a fgure that has been

fgures, based on triangulated data rom theInstitute o Applied Manpower Research, theAll India Council or Technical Education,

and NASSCOM, suggest that there was anestimated annual supply o 180,000 entry-level sotware proessionals (see Table 8),sufcient to meet current demand, consistentwith our interviews with HR managers whoargued that the shortages were in specifcskills at higher levels. (Compared to the previ-ously published estimates, ours are closer toestimates ound in a January 2008 study byGeref, Wadhwa, Rissing, and Ong.)

Other trends have been increasing turn-

over and declining average tenure: the per-centage o employees with less than two yearso experience increased rom 19% in 2003to 33% in 2004 (NASSCOM 2005). In 2003,77% o the organizations met their recruitingneeds through lateral hires (NASSCOM2004). Further, 32% o the 25,000 hiresplanned by Inosys in FY 20062007 werelateral hires. As a result, every sotware frmhas reported increases in turnover in recentyears (ranging rom 15% to 20%), leading to

a high degree o poaching at middle man-agement levels in the sotware segment.

The BPO segment This ast growing seg

Table 8. Indian Sotware Industry: Labor Supply Estimates.

Statistic 20032004 20042005 20052006 20062007

Number o Engineering Graduates 215,000 284,000 348,000 382,000Degree (4-year course) 112,000 155,000 210,000 235,000Degree (3-year course) 103,000 129,000 138,000 147,000

Number o IT (Computer Science, Electronics,Telecom) Proessionals 141,000 165,000 181,000 193,000

Engineering IT Graduates (degree) 95,000 100,000 111,000 117,000Engineering IT Graduates (diploma) 46,000 65,000 70,000 76,000

Number o IT Proessionals Entering the Work Force 80,000 94,000 103,000 109,000

Engineering IT Graduates (degree) 55,000 58,000 64,000 68,000Engineering IT Graduates (diploma) 25,000 36,000 39,000 41,000

Number o Non-IT Engineers Entering the IT

Work Force 40,000 40,000 40,000 Number o Graduates (Other Disciplines) Entering

the IT Work Force 30,000 30,000 30,000 30,000

Total Fresh IT Labor Supply 150,000 164,000 173,000 180,000

Source: Geref and Wadhwa (2006); NASSCOM (2006).


15/35


this industry is problematic, given the rapidgrowth rates (50% per year, on average),the increasing interest in outsourcing in the

West, and the expansion o Indias outsourc-ing ootprint.10 A NASSCOM-McKinseystudy projected a stock o 1 million jobs by20082009, while a NASSCOM-KPMG studyorecast a stock o 1.41 million by 2009. Tothe extent that past growth is an indicationo uture growth (at least over the next threeyears), the above estimates seem reasonable.Ater taking into account the current avail-able pool, the annual intake, productivityimprovements, and current rates o turnover,

the KPMG study estimates a shortage o262,000 people by 2009.11

We are aware o no readily available data,published or unpublished, regarding the sup-ply o English-speaking graduates (who areparticularly important or customer service jobs) to the BPO segment. Two problemsmake accurate orecasting difcult here. Thefrst problem is the absence o good statisticsregarding the output o graduates generally,12and the second, more critical problem is the

difculty in estimating the number o Englishspeakers available or the BPO call centers.Most human resource managers we talkedwith expected a much greater shortall thanthat estimated by NASSCOM,13 probably due

to dierential estimates o available English-speaking graduates. Table 9, which shows our

own estimates o HR supply or the low-endBPO segment, suggests that in 2005 therewas a pool o 250,000 graduates who werepotentially available or work. I we allow orvarious unaccountables, we obtain a moreconservative estimate o between 150,000and 200,000. But not all o these graduatesare capable o being hired. And given thatthe annual intake into the BPO segment isabout 120,000 and growing, a shortage o262,000 bodies (that is, o hirable workers,

quality aside) by 20082009 is a certainty.The high turnover rates (discussed below)are not entirely a unction o current short-ages, however.

Beyond the shortage o English-speakingworkers, our interview subjects highlightedshortages in industry-ready, industry-relevant manpower, that is, a scarcity oproessionals who are equipped with thenecessary domain knowledge to cater tospecifc verticals (industry segments) such

as banking, insurance, telecom, retail, andmanuacturing (Mitra 2004). Also remarked were shortages in the supply o oreignlanguage skills, notably in French, Spanish,Mandarin, German, and Italian, that make itdifcult or Indian BPOs to expand their pres-ence in European and Asian markets (IndianBPOs derive 82% o their business rom theUnited States and United Kingdom).

As o this writing (July 2008), there arerelatively ew current shortages in the R&D

segment. The fnancial segments need orskilled researchers and data analysts is be-ing met via the countrys large number omasters-level graduates in mathematics and

10One example o the difculty o making estimatesin this rapidly growing industry (and o using the past asa basis or uture predictions) can be seen in the experi-ence o recent years. In 20032004, the BPO industryadded 55,000 jobs. At that time, NASSCOM predictedthat over the next two to three years, the industry wouldgrow by about 55,00070,000 jobs. But in the next yearalone (20042005), 95,000 jobs were added.

11NASSCOMs estimate is that 1.416 million wouldbe required. Productivity increases would reduce thisfgure by about 413,000, leaving 1.003 million. Theexpected supply by 2009 is 741,000, leaving a projectedshortall o 262,000.

12The absence o fgures stems rom the division oresponsibility in the education sector o India. Thissector is on the concurrent list, which means that italls under the jurisdiction o the central governmentand state governments (Dossani 2006). In addition,

several states permit the private sector to operate schoolsand colleges. Furthermore, no central statistics agencycovers all these sectors.

to spoken English in educational institutions and privatetraining institutes. Second, it is possible that graduateswho have not been attracted to this industry will bemotivated to join it in the uture, especially given thegrowth in salaries. Finally, the number o educational

institutions may increase, consistent with prior trends,although it is unclear where the aculty will come rom(there is a shortage o aculty). In act, the number o


16/35


economics. In the lie sciences, there are no

shortages, as science postgraduates are avail-able currently. Neither are shortages reportedin engineering services, although the rapidgrowth in that arena will orce it to competewith the sotware segment or engineers, andwith traditional manuacturing, which is alsogrowing. In sotware R&D, there are, as noted,desperate shortages o visionaries. However,given the growth in the R&D services, moreskilled scientists and researchers will be neededquickly, and the Indian education system

cannot in the near term increase its outputo scientists and engineers (we discuss this ingreater detail arther below).

High Turnover

The industry is acing high rates o turn-over, ranging rom approximately 20% inthe sotware segment to about 5060% inthe BPO segment.

The software segment. In the case o the

sotware segment, average turnover was about20% during the boom years o the Y2K crisis(1999 2000) butsubsidedtonormal levels

dot-com bubble). While a range o 69%turnover in the top 10 frms was considerednormal until 2002 (interview, Limaye), in

July 2003 average turnover began increas-ing in these frms, reaching 22% by 2004, asIndian sotware frms stepped up their hir-ing. Wipro reported a turnover rate o 17%in 2004, and Inosys reported that turnoverincreased rom 7.9% in 2003 to 10.5% in2004 and 11% in 2005. Average turnoverin the top 10 sotware frms was over 20% inFY 20062007.

Attrition in the sotware segment variesby specialty and level. While most o the

HR managers we interviewed elt that theturnover situation is under control as long asthe attrition rates are high at the entry leveland not at the middle and senior levels, theentry and expansion o large MNCs suchas IBM, Oracle, and Accenture into Indiahave resulted in considerable poaching odevelopers, project managers, and high-levelexecutives. Currently (as o September 2008),attrition at the junior and middle manage-ment level is close to 3540%.

While we are unaware o any systematicstudies o turnover in the sotware segment,our sense is that a primary cause o attritionmore important than pay and beneftsis thelack o opportunities to work on cutting-edgetechnology. This is a young, very well-edu-cated work orce with high average earnings(see Table 10). A 2004 survey showed theaverage age o sotware proessionals to beabout 28 years; 76% were male, 50% hadmore than three years tenure, and 81% had

graduate degrees (69% were engineers, 17%had graduate qualifcations in engineering,cost accounting, chartered accounting, orbusiness, and 14% were diploma holdersor graduates in other social science felds)(NASSCOM 2005).

The BPO segment. Turnover estimatesrange more widely or the business processoutsourcing segment than or the sotwaresegment. Most reports o BPO turnover quotenumbers in the neighborhood o 3040%.

Commercial and popular reporting tendsto ocus on the extremes, that is, turnover aslow as 12% and as high as 90% One large

Table 9. Estimates o HR Supplyor Low-End BPO Segments, 2005.

Group Number

Output o Graduates (253universities and 13,150 colleges) 2,460,000a

Some English-Speaking Capability(30%) 738,000

Labor Force Participation (65%) 480,000b

Eective English-Speaking Abilityor Employability in Call Center (50%) 240,000

Eective Supply o Labor in 2005 168,000c

Source: Compiled by authors.aInstitute o Applied Manpower Research.bNASSCOM-KPM.cThis fgure appears to jibe with numbers obtained

rom our interviews with human resource managers(who, at the time we spoke to them, were claiming thatthe English-speaking market was already saturated). Thefgure is also consistent with another recent estimatethat 510% o the total graduate pool (123,000246,000people) speak English and are available to work in theindustry (Bhattacharjee 2003).


17/35


14,340 employees in its largely call center-o-cused BPO business in the last year (McCue2005:1). A 2004 study o turnover in 7 large

third-party call centers, which we considermore representative, showed average turn-over o 30%, with a range rom 12% to 62%(Remesh 2004). In general, turnover is lowerin captive BPOs than in third-party BPOs.We ound turnover rates in fve captive frmsto be between 1520%, although even somecaptives reported rates as high as 30%.

There is job-based variation in turnoveras well. Average turnover in voice-basedprocesses is 4550%, while in non-voice

operations it is 1520% (Remesh 2004).Further, inant mortality (turnover in thefrst 45 days o employment ater training) invoice-based processes is about 20% (interview,Varadarajan).

Causes of turnover. Although there havebeen no systematic published studies regard-ing the causes o turnover in the outsourcingindustry, our fndings, based on interviewswith HR managers14 who shared the results

o exit interviews with us, suggest that thereis remarkable consistency in the causes oturnover across frms. Most HR managers

we spoke to agreed, or example, that 50%o employees who leave a frm exit the indus-try completely, while 50% move to anothercompany. Most HR managers also seemedto identiy as the main causes o turnover theour we list below, although there was somedierence o opinion across companies aboutthe relative importance o each cause.

This consistency across HR managersresponses is not surprising, given that mostfrms are experiencing high turnover and that

BPO jobs and work orce characteristics aresimilar across BPO frms. The problem is soacute, yet diused, that some companies aresatisfed i they just succeed in maintaininga turnover rate o 20%, since this is belowthe industry average (interview, Ramesh).Even at conerences that bring industry HRmanagers togetherthe HR Summit, orexample, which the second author attendedin July 2007the ocus o the discussion is

Table 10. Indian IT Sotware & Services Industry: Median Pay.

Description ($US) Level 2004

Software SegmentSotware Engineer HA 1 6,560 6,040Senior Sotware Engineer HA 2 9,932 9,593Team Leader/Module Leader HA 3 14,820 13,190Project Leader HA 4 21,155 19,231Project Manager HA 5 30,816 26,653Program Manager/Sr. Project Manager HA 6 42,626 40,748HeadSotware Development/Large Business Unit HA 7 63,125 60,410

BPO Segment

Customer Care Executive (CCE) HA 1 (12 yrs.) 3,235 3,032Customer Care Executive (CCE)/Coordinator HA 1 (12 yrs.) 3,914Senior CCE/Executive HA 2 4,638 4,321Team Leader/Sr. Executive HA 3 7,240 6,855Assistant Manager HA 4 10,023 9,865Manager HA 5 17,240 17,987Assistant Vice President HA 6 36,110 33,056Vice President/Director HA 7 71,406 63,058

Source: Assembled rom NASSCOM Hewitt Total Rewards Study (2002, 2003, 2004).

14NASSCOM sponsors an annual summit or HR man-agers to discuss common HR problems. The frst author a sample o 12 HR managers rom predominantly large


18/35


not on addressing the underlying causes oturnover, but on the generation o peoplepractices to alleviate the problem (we discuss

these below). The same ocus is apparentin HR Connect, an online magazine or HRmanagers in the industry. Below, we examinethe primary reasons or turnover identifedby HR managers.

Job-related reasons. An important issue isdissatisaction with the immediate supervi-sor (the lousy boss syndrome). Becauseo the BPO segments youth (it is about 10years old), the youth o its new hires (mosto whom are 1927), and its high turnover,

promotion o young people with only 6months o BPO experience to positions inwhich they supervise teams o 1015 peopleis not uncommon. Relatively little time isspent training these people to be eectivesupervisors. These managers do not havethe required experience to manage teams oyoung people who themselves do not have ahistory o work experience, and who bring avariety o expectations and aspirations withthem when they enter the industry. HR

managers reer to this as the managerialbandwidth problem and liken it to the prob-lem o trying to create a 10-year-old scotchin 2 years (interview, Limaye).

A second key job-related problem lies inthe nature o the work itsel. There is consid-erable agreement in international research(Batt et al. 2006; Deery et al. 2004) that callcenter jobs are organized in a Tayloristicashion: they are highly repetitive, with tightlyregulated lunch and restroom breaks, targets

in terms o number o calls to be made, andintensive monitoring o employee activity.Indian call centers apparently ft this descrip-tion (Remesh 2004). Decrying the electronicsurveillance at work, a BPO worker said,Were treated as bonded labor (Nadeem2006:16). These workers, who are sometimespejoratively called techno-coolies,15 mustcope with considerable stress on the job (seeDCruz and Noronha 2006 or an elabora-tion). These workplace conditions and

othersespecially the night shit schedulesrequired to provide service to U.S. clientsduring North American daylight hourshave

resulted in an outbreak o what the industrycalls BOSS (burn-out stress syndrome), animportant cause o turnover. In response tothe growing number o employees seekingcounseling services and psychological help,more and more organizations (includingtrade unions) are oering such help. The Young Proessional Collective in Mumbai,or example, is organized by a well-knownlabor lawyer and staed by well-known laboractivists.

Finally, there are limited avenues or ca-reer advancement in a BPO frm (althoughthe labor shortage to some extent alleviatesthis problem). This lack o opportunity isborne out both by our own research and byprevious studies. Batt et al.s (2005) surveyo 60 call centers revealed that over the pe-riod o the study, 15% o the existing stocko employees at that time were promoted tohigher positions within the call centers, butonly 1% were promoted to upper manage-

ment levels beyond the call center sectionso the organizations. Our research suggestsa more nuanced picture: captive call cen-ters are able to provide more promotions tohigher levels than third-party call centers,because the market stability they enjoy os-ters longer career paths than in third-partycall centers.

Demographic profle o workers. Anotherimportant reason or turnover is the de-mographic profle o workers, raising the

question o whether the industry is target-ing the appropriate labor market segment.Most young people who work in call centersare marking time, with the goal o earninga good salary or a couple o years beorecontinuing their education. Second, callcenter employees are disproportionatelywomen, many o whom leave once they getmarried. The demographic picture o asubstantial cross-section o workers in a largethird-party BPO is particularly instructive.

Remeshs (2004) survey o 277 workers sug-gested that 75% o BPO workers were gradu-ates o convent schools (private schools that


19/35


or their superior and English-based educa-tion); urthermore, 94% o their athers and63% o their mothers, many o them working

or the government, were also graduates osuch schools. Members o this slice o theIndian middle class aspire to higher-levelproessional and government jobs, and areonly temporarily attracted to the higher BPOsalaries (the average BPO and call center sal-ary o Rs. 10,000 per month in 2005 representsabout 2.5 times what a graduate could earnin, or example, civil engineering).

Psycho-social factors. A number o psycho-social actors aect turnover. For upper

middle-class students, working on the nightshit is an alienating experience. They are notable to meet their riends in the evenings orto see their amily regularly. By the time theyget home, other amily members have goneto work. This distancing is new or theseyoung people. In other cases, middle-classparents o BPO employees disapprove o thework, complicating their childrens eelingstoward it. Pradhan and Abraham (2005)suggested that the new working habits and

patterns o living, that is, working late intothe night, the pub culture, the consumerism,are in complete contrast to ways o living to-gether, value systems, traditions and beliesstill nurtured by the middle class (p. 4).Other parents disapprove because they eelthat their children are lured by easy moneyand will orgo the chance to pursue highereducation. Yet others worry about theirdaughters reputation.16

Further, Pradhan and Abraham (2005)

have argued that having to adopt a dierentname and persona during work hoursanIndian girl named Anjali might becomeAngie at night, or example, and assume anAmerican accent when speaking to Americancustomerscan induce identity conusionand lead to multiple personality disorder,cultural sel-alienation, and a sense o dis-sonance. (See DCruz and Noronha 2006

or an argument regarding how the assumedname helps people deal with some workplaceissues.)

The actors described above are not theonly ones responsible or high turnover, butthey are the major reasons identifed by thehuman resource managers we spoke to. Asremarked earlier, although these respondentswere rom diverse frms in the BPO segmentand their comments did show some cross-frm variation, the level o their agreementregarding the main cause o high turnoverwas remarkable.

Increasing Human Resource Costs

Low cost has been a key advantage orIndias outsourcing industry. Table 11 com-pares costs in the United States and India ora number o jobs in both the sotware andBPO segments in 20022003. A triangulationo data rom various sources indicates thatthe cost o sotware and BPO employees inIndia was about 1520% o that in the UnitedStates. Rising costs are not a signifcant issue

in the highvalue-added portions o eithersotware or BPO work, where the competitiveadvantage is based more on skills. In sotware,cost increases have been incremental andsteady, by and large (one notable exceptionbeing a 1720% average increase in 2004).Until recently, these higher-end jobs tendedto be flled on a just-in-time basisthatis, hiring took place only when a companyobtained a contract or a projectbut theindustry has now shited to recruiting or the

bench, building up a stock o workers withskills that are in short supply, to be called onwhen needed. It is at the lower ends o bothsegments that the rise in costs threatens theindustrys growth.

Our research shows that about 45% othe average BPO organizations costs areaccounted or by employee expenses, 35%by telecom and bandwidth ees, and the re-maining 20% by other costs, such as admin-istration and marketing. The largest o those

categories, employee costs, has three criticalcomponents, all o which have been increas-ing: salaries; recruitment and training; and

16 An advertisement commonly used by critics oBPOs and their impact on traditional values in Indiagoes as ollows: Your daughter leaves home in the


20/35


sequence o the tight labor market and highturnover in this segment. Our research onlarger BPO rms suggests average total salarycost increases in the range o 1415% peryear over the 20032007 period, consistentwith the estimate o 12.5% by GE Capital In-ternational Services (GECIS), a leading BPOrm. The increase in compensation costs,however, has been almost perectly oset bya decline in bandwidth costs. Between 2001and 2005, average bandwidth prices ell byapproximately 20% per annum.17 Althoughbandwidth costs are expected to decline ur-ther, telecom costs have increased (largely asa result o the explosive growth o the BPOindustry).

However, it is the increase in the othertwo components o employee costs thatis the problem. The second componentconsists o recruitment and training costs(including training time). Given a declin-ing recruitment conversion actor that, as o2007, averaged only about 4% (only 4 out oevery 100 interviewed could be hired), rmshave had to cast larger and larger nets, and

the associated costs (advertising, recruiterstravel to other cities and university campuses)increased by over 25% between 2004 and2006, according to the managers we spoketo. As or training, it has grown more costlyin almost every respect, driven, in large part,by shortages o trainers. In the call centerindustry, it takes about 14 weeks o trainingbeore an employee is ready to work in a jobservicing the U.S. market. There are threetypes o training. The rst type includescommunication, grammar, vocabulary, pro-nunciation, and accent neutralization.The objective o such training is to ensurethat communication is ree o rst languageinfuence that impedes comprehension,oten reerred to as Indianisms in industryparlance (Ramachandran 2006:1). To quotea training manager,

Ater rigorous training in neutralizing accents,voice modulation, moderating the rate o speech,intonation and so on, call center associatesstep into their new personas. Thats where an

Alok transorms into an Alex. (InotechOnline2006:2)

The second type o training, which overlapswith the rst to some extent, is training in theclient countrys culture (loosely dened, andocused on certain icons such as baseball); in a

number o sot skills (ability to build rapportby using the appropriate tone, or example);andin comprehensionskills (understanding

Table 11. Per Hour Wage Dierentials between the United States and India, 20022003.

Silicon ValleyProfession U.S. Wage per Hour Indian Wage per Hour Wage per Hour

Telephone Operator $12.75 Less than $1.00 $13.24Health-Records Technologists,

Medical Transcriptionist $13.17 $1.50 to $2.00 $14.54Payroll Clerk $15.17 $1.50 to $2.00 $19.50Data Entry Clerk $20.00 $1.50 to $2.00 $24.44Legal Assistant, Paralegal $17.86 $6.00 to $8.00 NA Accountant $23.35 $6.00 to $10.00 $2Computer Programmer $28.90 $3.00 to $10.00 $38.85Financial Research Analyst $33.00 to $35.00 $6.00 to $15.00 $34.00Sotware Designer $60.00 $6.00 NASotware Engineer $120.00 $18.00 NAEntry-Level Programmers (annual salary) $50,000 to $60,000 $8,000 to $10,000 NA

Sources: Greene (2006)

17Although there is a lot o variation in the ways inwhich people calculate bandwidth pricing (the normalway is to use the cost o leasing a 1 Mbps link between two

cities, such as Mumbai and New York), most data show asteady decline, and this decline has more than oset theincrease in wage costs at BPOs or some time. GECIS


21/35


accents). Both o these components, accord-ing to a communications coach within a BPO,help call center agents zap the gap and earn

credibility (Ramachandran 2006:1).The nal type o training is in technical

and project-specic skills. This trainingis particular to the project or industry theemployee is associated with (or example,nancial products).

As the recruitment conversion actorsdecline, training time increases. Shortagesin training sta, especially in sta who candeliver instruction in culture and accentneutralization, contribute to increasing salary

costs. Arguably, these costs would decreasei the activities could be outsourced, but theHR managers we interviewed in this industrysaid they were not satised with the qualityo labor market intermediaries (rms thatwould recruit, train, and supply sta).

The nal component o employee costsconcerns employee transportation to androm work (a cost that in other countrieswould be borne by the taxpayer in the ormo ecient public transportation systems).

Given the absence o eective and sae publictransportation, especially at night (which iswhen call center employees work to servicethe U.S. market by day), BPO companiesprovide transportation to their workplaces(most are located on the outskirts o majormetropolitan areas). Saety concerns are soimportant (especially the saety o emaleemployees, who comprise about 50% o thework orce) that even i the most economi-cal route or a van carrying six employees

might be to drop o a emale person last,she must be dropped o beore at least onemale colleague. Training and discipline odrivers is also important. To keep a check onsecurity, a BPO company called EXL, based inNew Delhi, installs radio systems with whichdrivers can report back to the base stationevery 5 kilometers. Further, at least oncea week, EXL checks each company shuttleat a random time and place in the city (Ag-garwal 2003). The cost o managing a large

transport operation is high. Our estimates,based on interviews with HR managers, sug-gest that transportation costs account or as

oversee transport operations. For example,in 2004, EXL had a work orce o 2,600 and afeet o 150 vehicles rom 7 dierent vendors

to cover about 440,000 person kilometerseach day. The expenses per employee o thisoperation worked out to about Rs. 2,500 permonth, which is approximately a quarter othe average monthly wage. The industry aver-age was much higher, at about Rs. 3,500 permonth in 2004 (interview, Vijayabhasker).

Thus ar, increases in human resourcecosts have been partially oset by decliningbandwidth costs. And relative costs are notrelevant or the smaller part o the segment

that is engaged in higher-value-added activity.However, since low-cost customer care workstill constitutes the bulk o the industry, publicpolicy initiatives that could result in betterurban transportation systems or producemore English-speaking graduates could sig-nicantly reduce HR costs and help maintainthe industrys competitive edge.

Developing Long-Term High-

Skill Human Resource CapabilityO the our problems, this is perhaps the

most critical one, given the upward trajec-tory o the outsourcing industry. We arguethat Indias postgraduate education and re-search system cannot yet generate high-levelskills o the kind that are required or growthin the high end o the industry, particularlythe sotware and R&D segments. Severalindicators support this argument.

First, government spending on research

and development is considerably lower inIndia (0.0843% o the countrys GDP in 2007)than in some competing countries (2.8% oSouth Koreas GDP, or example). In act,government spending on higher education asa proportion o GDP ell rom 1% in 1971, to0.8% in 1981, to 0.6% in 1991; in 2001, it was0.4% (Mahadevan 2006). Further, the pro-portion o expenditure on higher educationto total expenditure on education declinedrom an average o 15% in the 1980s to 10%

in the 1990s (Kapur and Mehta 2004).The lack o ocus on research is refected

in patent data Table 12 shows the move-


22/35


1980 and 2005. While Korea increased itsshare rom 0.03% o all patents in 1980 to4.24% in 2005, India showed only a modestincrease, rom 0.021% to 0.37%. This is alsoconsiderably lower than the increase postedby China. Within India itsel, more oreign-ers than Indians have been fling patents:in 20032004, oreigners obtained 1,400patents, Indians 1,200. India also ranks verylow on the technology achievement index

(0.21, compared to the United States 0.73,in 2001), and has one o the lowest numberso research and development personnel percapita (World Bank 2001).

Second, India is not appreciably increasingits production o doctorates in scientifc felds(critical or the uture o the outsourcingindustry). The total number o doctoratesin science produced in India has been stag-nant (rom 3,861 in 1996 to 3,896 in 2000). What is even more worrisome is that the

proportion o doctorates in science out ototal doctorates has been steadily decreasing(rom 50% in 1974 to 38% in 2002), whilethe shares o doctorates in arts and businesshave increased. Nor has the supply base opotential doctoral students been growing;on the contrary, there has been a long-termdecline in the proportion o students who arestudying science in college, which ell rom31% in 1950 to 20% in 1990 (University GrantsCommission 2007). Further, the number o

engineering and technology Ph.D.s receivedin India is low (298 in 1997)indeed, substan-tially lower even than the number o Ph D s

Clearly, ew Indians wishing to undertakescientifc research choose to do so in thedoctoral programs o Indian universities.Most o them instead preer either to entereducational institutions outside India, wherethey can more easily work on cutting-edgeresearch, or to join industry research stas,where they are paid a higher salary than theycould obtain itaking the other main poten-tial career path open to them in Indiathey

chose to become research assistants or associ-ates (whose pay is low because it is capped bythe University Grants Commission).

Third, the poor quality o doctoral edu-cation in India has been a long-standingproblem (Chatterjea 2007). The India Sci-ence Report published in 2005 ound that1415% o doctorates in science were notemployable ater their studies, in spite o therapidly increasing demand or researchers.Chatterjea and Moulik (2007) argued that

although doctoral education in colonial In-dia produced a number o world-renownedscholars, doctoral education since indepen-dence (with a ew notable exceptions) hasbeen hampered by poor aculty quality, alack o resources, poor leadership, and severeinrastructure problems.

It is not hard to fnd concrete evidence thatIndias advanced education system suers(with a ew notable exceptions) rom qual-ity problems. For example, in 2007, not a

single Indian university was among either theworlds 300 top-ranked universities (as com-piled by the Institute or Higher Education at

Table 12. International Utility Patents Filed in theUnited States, 1980, 1990, 2000, and 2005, Selected Countries.

Country 1980 1990 2000 2005

U.S.A. 62,098 90,643 164,795 207,867 Japan 12,951 34,113 52,891 7S. Korea 33 775 5,705 17,217Taiwan 367 2,035 9,046 16,617Brazil 53 88 220 295China 7 111 469 2,127India 23 58 438 1,463

World Total 104,329 164,558 295,926 390,733

Source: U.S. Patent and Trademark Ofce


23/35


o excellence such as the Indian Instituteso Technology [IIT]). India occupied the33rd position among all countries in terms o

research quality (the United States ranked1st, China 19th, and South Korea 21st).

It is thereore not surprising that Indiadoes not are well in terms o the quality oits research either. Over the 19982005 pe-riod, Indias shareo all papers published injournals listed by the Science Citation Indexdid increase by 14% (rom 1.57 to 1.79), butChinas increased 93% (rom 1.90 to 3.68),and Brazils, 44%. Further, even in the bestinstitutions, citations per aculty member

are low; at the IITs, or example, the 5-yearcitation count per aculty member has been2 or 3, compared to 45 at MIT and 52 inStanords engineering department (Hearne1999). Thus, advanced educational institu-tions in India have historically done little basicresearch, and such research as they have donehas not spawned business, as has happenedin Silicon Valley. As one proessor noted,We dont have an entrepreneurial culture,nor have we a well-developed research inra-

structure (interview, Parthasarathy).The poor quality o advanced education

is urther refected in engineering (whichis critical to the nations economic develop-ment). Although there has been an increasein the number o privately run engineer-ing colleges in many Indian states, whichhas contributed to the increasing outputo engineers (Puliyenthuruthel 2006), the All India Council o Technical Educationhas eliminated about 22,722 seats at 300

engineering colleges because the collegeslacked aculty, courses, and inrastructure.O course, the countrys rst-tier and second-tier engineering institutions, such as the IITsand the Regional Engineering Colleges, areworld-amous or their student quality, butit is the third-tier engineering colleges thatproduce 88% o engineering undergradu-ates, 66% o engineering postgraduates,and 82% o engineering aculty (Cairncross2006). Some o these colleges have neither

electricity nor access to computers, suerrom a shortage o aculty, use textbooks thatare decades old emphasize rote learning

A key issue is the shortage o aculty. Iengineering colleges have increased in num-ber, engineering Ph.D.s have not, raising

the question o where the aculty in thesenew colleges are coming rom. Even in thebest institutions, such as the IITs, there arecritical shortages. Several departments at IITDelhi are acing a 40% shortage in aculty,the problem being most acute in computerscience, chemical engineering, and electricalengineering (Sinha 2002). The other IITs(IIT Kanpur, IIT Kharagpur, IIT Bombay, IITMadras, IIT Guwahati, and IIT Roorkee) areacing similar shortages in aculty, between

20% and 35%. For example, Sinha noted thatwhen IIT Delhi advertised aculty vacanciesin its chemical engineering department, 120applications were received, o which 18 wereselected or a nal interviewbut not oneo the interviewees was deemed qualied orthe positions. The aculty shortage is evenmore acute outside the engineering arena:in 2006, at non-engineering departments inregular public and government-aided col-leges in ten Indian states, more than 22,500

vacancies or lecturers went unlled or lacko applications.

It is not our intention to recapitulatethe considerable prior literature that hashighlighted problems acing Indias highereducation system in general (see Chatterjeaand Moulik 2007). But even our selectivereview makes it quite clear that the countryshigher education system may not be equal toturning out highly educated engineers andresearchers either in the numbers or o the

quality required to meet the uture needs oa growing outsourcing sector.

Summary

The HR problems we have discussed donot uniormly aect the dierent segmentso the outsourcing industry. Skill shortagesaect both sotware and BPO segments.The problem o high turnover is particularlysevere or the BPO segment, less so or the

sotware segment, and, at least as o this writ-ing, non-existent or the R&D segment. Theproblemorapidly increasingHRcostsaects


24/35


three segments, but especially the sotwareand R&D segments. The varying severity othese problems across the dierent segments

makes it difcult to determine precisely howthey are aecting the growth o the wholeoutsourcing industry. In addition, the highgrowth rate o the industry in general makesaddressing these problems less urgent romthe standpoint o policymakers and frms.

However, industry leaders have increas-ingly been complaining about lost businessopportunities. The high turnover has ledto increased poaching. Some clients havemoved India-based call centers to the Philip-

pines and China. Some sotware outsourcingopportunities have been lost to Romania, Hun-gary, and Estonia. Perhaps most compellingas a sign that skill shortages are aecting busi-ness is the requently heard complaint aboutthe quality o new recruits. One manager ina well-known Wall Street frm that has its ownsotware support unit in India told us that weare really scraping the bottom o the barrelnow and are hiring engineers rom third andourth tier schools who have very little engi-

neering knowledge at all (interview, Rajan).Another manager told us in June 2007 thatin his opinion, one should expect to see thegrowth o the Indian industry begin to decline,given the declining quality; The boom is bust,he said (interview, Krishnakumar). On theother hand, India still remains the dominantdestination or a variety o outsourcing workthat requires English-speaking ability. Thus,the current high growth rates mask to somedegree the immediate eects o underlying

skills shortages, but the long-term sustainabil-ity o the industry is surely in question i someo the macro problems are not resolved.

Addressing Macro and Micro HR Problems

The combined problems o skill mismatch-es, skill shortages, and hindrances to long-term skill development are not unique to theIndian outsourcing industry. Such problemstypically arise when a new industry develops,

or when there is a change in developmentstrategy. For example, high turnover was akey problem in the early development o the

Okayama 1986), and the U.S. automobileindustry (Sward 1968). Similarly, Singapore,Korea, Taiwan, and Malaysia experienced skill

shortages and skill mismatches when theytransitioned rom a low-cost labor-intensiveEOI strategy to a highervalue-added EOIstrategy (Green et al. 1999; Kuruvilla, Erick-son, and Hwang 2002).

Solutions to such problems have diereddepending on the institutional context. Inthe silk industry in Japan, frms introducedthe concept o lietime employment to solvethe high turnover problem (Taira 1970). Inthe Japanese auto industry, employers took a

long-term view to solve skill shortages, highturnover, and labor management relations bydeveloping an internal labor market involvinglietime employment, frm-specifc broad-based training, and enterprise unionismin-stitutions that are now well established asdistinguishing eatures o the Japanese indus-trial relations system. In the case o the U.S.automobile industry, Henry Fords responseto labor strie, absenteeism, and turnover wasto introduce the 5-dollar day; subsequently,

Sward (1968) wrote, the companys laborturnover dropped by 90% [and] absentee-ism rom 10% to 1/3rd o 1 percent (p. 56).To minimize skill mismatches and optimizelong-term skill development upon transitionto a new industrialization strategy, Singa-pore adopted a concerted national eortthat involved multiple actors, encompassingmultiple institutions, at national, university,and frm levels through a variety o policymechanisms, include taxation (Kuruvilla,

Erickson, and Hwang 2002). Singaporesapproach and, more recently, Malaysias areconsistent with Singh and Zammits (2000)argument that avoiding skill mismatches andupgrading skills or economic developmentrequire a concerted multi-level national e-ort, with multiple mechanisms and policiesthat involve multiple actors. Solving bothmacro and micro HR problems in Indiasoutsourcing industryproblems requiringdierent policy interventions by dierent

actorsimplies, similarly, a concerted, orwell-integrated, national eort.

Finegold and Soskices skills equilibrium


25/35


concrete policy advice, and to evaluate thecurrent approaches in India. Finegold andSoskice (1988) argued that moving rom one

level o skill equilibria to another (avoidingskill mismatches, in other words) requiresthat dierent actors in the system take con-certed and reinorcing action. For example,individuals must be willing to invest in appro-priate education and training, rms must bemotivated to develop and retain employees,and governments must be motivated to en-sure that there are no skill shortages. Threecategories o institutional and market-basedconditions are prerequisites or orcing this

kind o action: conditions that orce actorsto take a long-term outlook to ensure thatinstitutions counter pressures rom capitalmarkets to ocus on the short-term; actorsthat encourage inter-rm cooperation withina

Documents

Out Sour Sing India