Click here for slides

(c) Ashish Arora, 2006, Toronto Offshoring Conference

The Indian software industry and its prospects

Ashish AroraHeinz School, Carnegie Mellon University

The Software Industry Center at Carnegie Mellon University, and the Alfred P. Sloan Foundation provided support for this research.


Outline

The software industry – an activity, not an industryHow did the underdog become a tiger?Prospects: Moving up the value chainImpact


Outline

The software industry – an activity, not an industry

Most software workers do not work for IT firmsThe software activities inside non-IT firms were main target of Indian SW exports

How did the underdog become a tiger?Prospects: Moving up the value chainImpact


Software is Ubiquitous: SW Employment as % of Total Employment by State, 2001

0-1%1-2%2-3%>3%

Source: BLS, 2001


66% of US software jobs are outside the IT industry

Industry SW Employment

Computer Equipment 72,000

Computer and Software Services

977,000

All Other Industries 2,816,000

Total 3,865,000

Source: BLS, 2001

Software is both a technology and an industry


SW is both an activity and an industry, and US government statistics have not caught up

1. Salaries of temporary workers (~40%)2. Embedded SW (~10-15%)3. Sales to US MNC affiliate overseas4. US counts only unaffiliated imports

Source: GAO, Oct 2005


($ billion)2002-03

2003-04

Banking, Financial Services & Insurance (BFSI)

39% 40%

Manufacturing 12% 12%

Telecom equipment

9% 9%

Government 1% 1%

Retail 5% 5%

Utilities 3% 3%

Transportation 1% 1%

Telecom service providers

4% 4%

Healthcare 5% 5%

Others 21% 21%

Total 100% 100%

Sou

rce:

ww

w.n

asscom

.org

Indian SW exports target the in-house SW activities of user firms: Offer business services

Standardization and modularization of “support” functions inside large organization, which reduced (but not eliminated) the need for local presence

Demand-Supply mismatch in 1990sTelecommunication advances

Much of globalization is directed at “in-house” software systems, through outsourcing, increasingly through longer term partnershipsUS vendors are starting to increase offshore capabilitiesUS firms are starting offices and subsidiaries to provide business and IT service support

Indian IT software services and BPO (ITES) revenues: Vertical


Outline

The software industry – an activity, not an industry

How did the underdog become a tiger?

Human capital, bad economy, good luckBenign neglect – policy played a minor roleBangalore ?

Prospects: Moving up the value chainImpact


The SW industry in India has grown rapidly ~ 30% p.a. since 2001.

Employment in the Indian Software and Services sector

Sector 2004 2005 2006

IT Services 614,000 741,000 878,000

ITES-BPO 253,000 316,000 415,000

Indian IT Industry-Sector-wise break-up

USD billion FY 2004 FY 2005 FY 2006E

IT Services 10.4 13.5 17.5

-Exports 7.3 10.0 13.2

-Domestic 3.1 3.5 4.3

ITES-BPO 3.4 5.2 7.2

-Exports 3.1 4.6 6.3

-Domestic 0.3 0.6 0.9

Engineering Services and R&D, Software Products

2.9 3.9 4.8

-Exports 2.5 3.1 3.9

-Domestic 0.4 0.7 0.9

Total Software and Services RevenuesOf which, exports are

16.7 22.6 29.5

12.9 17.7 23.4

Hardware 5.0 5.9 6.9

Total IT Industry (including Hardware)

21.6 28.4 36.3

Source: Nasscom (IT factsheet), www.nasscom.org (accessed 18 Sept 2006)

Chinese SW exports ~$3.5 BProjected to increase ~12 B

by 2012


How did the underdogs turn into tigers?

Economic Development with Unlimited

Supplies of LaborAbundant human capital supply

relative to domestic needIreland and Israel had poor growth in 1980s; Indian economy stagnated in 1970sSupply elasticity high

Openness and links with export markets

Falling telecom costsExpatriatesEnglish languageMNC contacts (esp Ireland)

Good timing and luckliberalization when economic boom and global IT skill shortage

Firm levelStrong entrepreneurial response and accommodating policy

High rates of entryexperimentation

• market (geography)• market (product or service)• business model (e.g.,

service delivery)

learning and capability acquisition

Lack of organizational capability perhaps why other English speaking, human capital abundant countries such as Philippines, Pakistan and Bangladesh did not make it.


Human capital is key

KA

TNMH

AP

DL

HA

UP

WB

OA

KL

MP GJPN

RJ

68

10

12

Log (

Softw

are

Export

2003 -

Softw

are

Export

s 1

990)

(Rs B

N o

f 1993-9

4)

1957 1984 1983 1977 1999 1995 1995 1996 1986 1992 1986 1995 1993 1998Year of Policy Change

Entry of first private engineering college

•Up to late ‘1970s most of colleges in public sector •Now very large share of private self-financed institutions•Undergraduate engineering capacity has grown seven-fold between 1990 and 2003 •Large inter-state variation in intake capacity•Vast bulk of capacity differences is due to private, non-funded, colleges

SW

export

s 2

00

3 –

SW

export

s 1

99

0


Difference in state education capacity predates the rise of SW exports.

State Share of Private Non Granted College in Sanctioned Engineering Baccalaureate Capability

0%

20%

40%

60%

80%

100%

120%

WB OA UP HR PN DL RJ MP AP TN KA KL MH GJ

STATES

% N

on-G

rant

ed E

ng. B

acca

laur

eate

Cap

acity

198719952002

Source: My calculations based on AICTE data on sanctioned capacity

Avg. Engineering College Capacity

Early Adopters

Late Adopters

1991 9,258 1,889

2003 42,144 12647

Diff 32,886 10,758

Avg. Software Exports

in millions of Rupees, 1993-94 prices

1991 602 102

2003 39,682 9,874

Diff 39,080 9,772Source: Arora & Bagde, 2006


Eng. College Capacity 1987

5.96(1.00)

Electronics Production 1990

0.97(0.50)

Lagged Industrial Output 1987

-0.56(0.15)

Constant 6096(4956)

R2 0.90

No. of obs. 14.

States that allowed private engineering colleges early likely to emerge as SW hubs

Eng. college capacity (-4 yr) 0.34 (0.1)

0.20 (0.07)

electronics production (-1 yr) 0.40 (0.24)

industrial output (-1 yr) 0.007 (0.023)

per capita income (-1 yr) -0.55 (0.61)

Population -0.28 (0.16)

Constant -371(1308)

22981(11914)

State fixed effects Yes Yes

Year effects Yes Yes

R2 0.49 0.54

Note: Cluster corrected std. errors in parenthesis. No. of obs. 182.

Dependent variable: Annual change in SW exports

Dependent variable: SW exports 2003 – SW exports 1990

Source: Arora and Bagde, 2006


SW exports OLS

SW exports 2SLS

SW exports 2SLS (2)

Eng. College Capacity (-4)

0.20 (0.07)

0.62(0.36)

0.74(0.50)

Electronics Production (-1)

0.40 (0.24)

0.21(0.23)

Industrial Output (-1)

0.007 (0.023)

-0.03(0.05)

Per Capita Income (-1)

-0.55 (0.61)

-0.67(0.67)

Population (-1)

-0.28 (0.16)

-0.15(0.14)

Constant 22981(11914)

-4773(4489)

9397(11527)

State-fixed effects Yes Yes Yes

Year-fixed effects Yes Yes Yes

R2 0.54 0.45 0.44

Note: Cluster corrected std. errors in parenthesis. No. of obs. 182.

Results survive controlling for reverse causality

Instrument for eng. college capacityMean of neighboring states’ education policy

education policy for a state is dummy variable = 1 when first self-financing college starts and stays 1 thereafterIn 1991 only 6 out of 14 states had self-financing colleges By 1998, all 14 states allow

Shows the benefits of decentralization


Public Policy?

Liberalization very helpfulFiscal other benefits usefulHuman capital investments were not intended to foster SWComparative advantage as much a result of poor economic growth as high human capital

Affirmative, sector specific policies unimportant

Policy changes inconsistent

SW flew under the radarNaturally exempted from many labor and industrial regulations

Hardware, the focus of much policy attention (till SW succeeded!) has very little to show


Bangalore was not the past of the Indian SW industry, though it may be the future

Share in SW revenues, major states, 1990-2003

0%

10%

20%

30%

40%

50%

60%

1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003

KA TN MH AP DL LocationPre-1980

1981-84

1985-91

1992-99

2000 -01

Bangalore 3 3 19 50 15

Mumbai/Pune(Pune)

9(1)

11(0)

32(8)

63(17)

8(2)

Chennai 3 5 9 34 6

Delhi: of which(Noida)(Gurgaon)

5 4(1)

25(6)(1)

63(18)(9)

17(4)(2)

Hyderabad /Secundrabad

1 6 29 8

Entry dates and the regional location of firms, 2001

Source: Arora and Bagde, 2006

Source: Athreye, 2005

Little merit in claims that Public Sector R&D labs (in Bangalore) explain the growth of the Indian SW industry


Learning and the domestic market?

Not importantExports were different in nature – technical sophistication was of limited valueBrazil has a very sophisticated domestic banking and telecom sector, served by domestic software industry but very little by way of exports.HW had ample protection and very little success

Israel has been more successful in network securityi-flex did succeed in leveraging domestic experience for exports

“… (Our parent firm) … was the first firm to use IBM mainframes in India for a very long time … We have the most qualified experts on IBM mainframes. … (But) technology is not such a critical factor as compared to understanding business practices.”… Domestic expertise may be useful in gaining technical expertise such as in coding and project management. However domestic and export projects are two different ball games.”(Interviewed by the author in Bombay, 1997, quote extracted from Arora et al., 2001. Emphasis added.).


Entrepreneurship and openness

“When I was out there in 1991, the country was bankrupt. We had three governments in one year, an assassination of a prime minister, and we were hawking our gold. You know, selling overseas was not a piece of cake…. if I have to present ten slides, the first eight had to be to sell India and the ninth one would say we do have an IT industry in India and unless the guy bought those nine slides, your tenth one about your company was meaningless. Because who are you anyway? Fifty people -- its no big deal. So we were building up the (India) brand from day one”

(A founder-member of NASSCOM, interviewed by Suma Athreye, 2005, cited in Athreye and Hobday, 2006)

Comparative advantages is very broad.Figuring out what will work under the given institutional realities is a hugely creative act (Hausman and Rodrik)

Product –service; prime-subcontracts; offshore-onsite; …How to execute with 40% turnover of employees, capital constraints, poor brand image outside, need for close client interaction …

That there are successful Indian SW firms is noteworthy.

Otherwise Saudi Arabia should have the leading oil firms and UK should have had the leading organic chemical firms in the 19th century

Indian SW firms have, after considerable experimentation and effort, developed the hybrid delivery model which uses talented but poorly trained and inexperienced workers.


Openness and Entrepreneurship

Name of firm Year Est. Origin/type of firm Notes

TCS 1968 Business house Founder US educated

Wipro 1980 Business house

Infosys 1981 Spin-off (Patni)

Satyam 1987 Business house Founder US educated

HCL 1991 Entrepreneur

Patni 1978 Entrepreneur Diaspora

I-flex 1989 Spawn (Citibank) MNC spawned

Tech Mahindra 1988 Business house

Perot Systems 1996 MNC (earlier joint venture with HCL)

L&T Infotech 1996 Business House

Polaris 1993 Entrepreneurial

Hexaware 1989 Entrepreneurial (Venture funded)

Mastek 1982 Entrepreneurial

Mphasis BFL 1992 Spin-off (Citibank) Diaspora

Siemens MNC

Genpact 1997 Spawn (GE) MNC - Diaspora

IGate 1993 Entrepreneur (US based) Diaspora

Flextronics 1991 MNC (Hughes Software) - Diaspora

NIIT 1981 Entrepreneur HCL spawned

Covansys India 1985 Entrepreneur (US based) Diaspora (CBSL)

PLUSIBM, •Accenture•HP•Syntell•Inteligroup•Kanbay

•Ireland: 75% firms had a founder who had worked broad (Sands, 2005)•Israel: 40% of managers of listed firms had US degrees•India: 1/3 firms by MNC or diaspora

NASSCOM Top 20 SW Exporters

•Diaspora brokers supply and demand•TI and Citi pioneer offshore model


Outline

The software industry – an activity, not an industryHow did the underdog become a tiger?

Prospects: Moving up the value chainFor established firms: Move up the value chain but not the technology ladderFor India as a place to develop software: Move up the technology ladder

• Products• R&D• Engineering Services• Threat of automation: SOA, SW as a service …

Impact


U.S. continues to lead in SW research and innovation

Source: ACM Report on Globalization and Offshoring of Software

P

Presentation made to the STEP Board of the National Academy of Sciences, Arora, Forman and Yoon

G-7 Software Patenting, US PTO filings

0500

100015002000250030003500400045005000

Years

Numb

er of

US

paten

ts

Canada France German Italy Japan United Kingdom USA


Israel, Korea and Taiwan lead among emerging

"Underdog" Software Patenting

0

10

20

30

40

50

60

70

80

90

100

Years

Nu

mb

er

of

US

pate

nts

Brazil India China Ireland Israel

Four East Asia Dragons

0

20

40

60

80

100

120

140

160

Years

Nu

mb

er

of

US

pate

nts

Korea Taiwan Singapore Hong Kong

Most Indian invented patents are assigned to US firms but S. Korean and Taiwanese firms for patents invented in those countries

MNCs are organizing SW innovation in India but not in the Asian Tigers

Source: USPTO and author calculations.

Presentation made to the STEP Board of the National Academy of Sciences, Arora, Forman and Yoon


Patenting Activities of Leading Software Firms in India, USPTO patents invented in India by type of assignee

Company2004-05 2005-06

Filed

Granted

Filed Granted

Microsoft 40 - 70 -

Symantec 47 43 57 16

ST Micro 62 32 37 14

Adobe 10 - 32* -

Freescale 10 - 16 4

Flextronics 2 1 4 1

Cadence 1 5 -- -

Texas Instrument

35 10 - -

Panel B: Indian SW patenting firms

Company2004-05 2005-06

Filed Granted

Filed Granted

Infosys - - 20 -

Ramco 16 - 16 -

TCS 16 5 13 4

Sasken 5 - 5 5

Mindtree 1 - 2 -

Subex - - 2 -

Source: DQ estimates All figures represent US patents filed by the Indian R&D facilities of these MNCs in all fields, not just software patents*Adobe's patents represent the total number of patents filed since start of India operations

Panel A: MNCs patents invented in India

Inventive activity from India is growing but dominated by MNCs, and in areas such as semiconductors


Summary of evidence on innovation and R&D

Current evidence suggests that North America continues to play to predominant role in software innovationA key reason for the US dominance is that the US is home to most of the lead users of SW

Most SW is used to run businesses, not to do technical things.Being a talent magnet also helps

India will be a good place to locate some R&D functions, particularly those that do not need tight integration with other user facing functions

• Most still from multinational firms

• More upstream technology – semiconductor, device driver, embedded SW

2.9 3.9 4.8

-Exports 2.5 3.1 3.9

-Domestic 0.4 0.7 0.9

USD billion FY 2004

FY 2005

FY 2006E

Company Revenue (2005-06) $ million

Growth (%)

HCL Technologies

222 40

TCS 196 62

Satyam 82 53

Rolta India 31 30

Quest 15 40

Neilsoft 8 40

Leading Indian Engineering and R&D Service Companies

R&D, Engineering Services and SW Products Revenues


Academics and Industry leaders united in this opinion

'Focus on IT Products, Not Outsourcing‘

Indian software entrepreneurs would need to focus more on innovation of new IT products rather than on services or outsourcing if India had to be at the forefront of the IT revolution and transform itself into a software powerhouse, according to founder of Hotmail, Sabeer Bhatia. (15 July 2005, Bangalore, The Economic Times)

Onsite working increases the opportunities for a ‘brain drain’ of talent,

… and may also leave the Indian industry unable to move significantly to

a different form of exports, such as package exports.

(Heeks, 1998, India's Uneven Software Exports, Working Paper, IDPM, Univ of Manchester )

Is products the way to move up the value chain?

2004-05 2005-06

i-flex 122.3 157.7

Infosys 44.4 74.2

TCS 41.0 53.5

3i Infotech 26.9 40.6

Cranes 27.9 34.2

Ramco 25.8 32.9

Tally 47.7 28.5

Subex 13.1 24.4

Flextronics 14.8 23.8

Polaris 24.6 15.4

Total 388.5 485.2

Leading software product companies and software product revenues, $ million.

Leading Indian software firms earn less than 2% from software products


Percentage of Total 2002 Software Product Exports and Imports by OECD Country

0% 5% 10% 15% 20% 25%

AustriaCanadaFrance

GermanyIreland

ItalyJapanKorea

NetherlandsSweden

SwitzerlandUnited

United StatesAll Others

Exports % Total Imports % TotalSource: OECD

Source: OECD (2002) using IDC data. Reported in Thoma and Torrisi (2006).

Packaged Software Sales by Region, 1990-2001 (US Dollars)

The North American Dominance of the

SW Product Market is Remarkable

Over 35% of exports and over 50% in sales!

Substantial high end human capital moving to the US.

Large fraction of engineering grad students from overseas.

US remains major market. New product development needs to be close to market and to lead users

Institutions for supporting innovationSuccessful product innovators from India,

Ireland, Israel … move to the US (Talisma, Checkpoint, Baltimore)


REVENUES 2004 2003

New license 3,541 3,270

License updates 4,529 3,929

Software 8,070 7,199

Services 2,086 2,276

Total rev 10,156 9,475

Operating Expenses

Sales & Market 2,136 2,072

License updates and prod support

547 474

Cost of services 1,770 1,868

R&D 1,278 1,180

G&A 561 441

Total Expenses 6,292 6,035

NET INCOME 2,681 2,307

ORACLE : Income statement

REVENUES : 2004 2005

Consultancy services 1461 2023

Sale of equipment and software licenses

108 118

Other revenues 13 20

Total revenues 1583 2162

Operating expenses:

Selling, general and administrative expenses

314 443

Cost of services 760 1087

Cost of equipment and software licenses

97 111

R&D 7 7

Net income 405 513

TCS Income Statement

Very similar overall margins (24% v 21%)! Products are not a panacea.


Outline

The software industry – an activity, not an industryHow did the underdog become a tiger?

Human capital, bad economy, good luckBenign neglect – policy played a minor roleBangalore may be the future, not the past of Indian software exports

Prospects: Moving up the value chainFor established firms: Move up the value chain but not the technology ladderFor India as a place to develop software: Move up the technology ladder

• Products• R&D• Engineering Servicesl

ImpactNow and futureEmployment?Intangibles: catalyst, exemplar and legitimizer


Examples of large contracts obtained by Indian SW firms

DateIndian firm Client Contract type Value (million)

( period)

2006 Wipro GM $27-$300

Sep 2005 TCS ABN SW Dev $260

Sep 2005 Infosys ABN SW Dev, maintenance $140

Aug 2003 L&T Motorola $70-90 (3-5 yrs)

Aug 2003 Satyam Certain Teed (USA) Implement supply chain solution. $15 (9 months)

Jun 2003 HCL Airbus Embedded SW -

April 2003 HCL B T group (UK) Business telemarketing, billing conferencing

$ 160 (5 years)

April 2003 Infosys BT group (UK) Second service provider for BPO services

- ( 5years)

Mar 2003 Patni r Guardian Life (US) Gap analysis and implementation. $35 ( 7 years)

Mar 2003 Ramco- Boeing

Aloha Airlines (US) Technical services with main marketing by Boeing (50% of revenues for each )

-

Nov 2002 TCS & Wipro

Lehmann Bros. IT outsourcing $50-70

Jan 2002 TCS GE medical ‘Take or pay’ model, $100-120 ( 2 yrs)

July 2001 Wipro Lattice Group ( US) Outsourcing $70 (3 years)

Leading Indian firms are moving up the value chain, but not necessarily the technology ladder, AND this is OK.


Prospects : Rosy (or at least not bleak)

Established firms move up the value chain but not the technology ladder

Become more like EDS or AccentureNo Microsofts or Oracles or Googles in the offing – but that is nothing to be ashamed of either.

Product development will be controlled by US firms

Captive developmentOutsourcing

Smaller and newer firms will try for innovation and products

Threats?Supply of human capital?

Talent is plentiful, skills can be scarceProbably not a very serious constraint because very responsive supply side

Urban infrastructure?Will likely be available

Competition from US incumbents?Matters for firms, not country

Automation: Rise of service oriented architectures (SOA), virutalized applications, utility computing …

Do not host the software and therefore do not maintain it and upgrade itInstead rely upon infrastructure providers such as IBM, HP and Oracle to offer the services of the software applicationsNo need for in-house software workers NOR for Indian software firms.

Lack of hardware and integration capability will hurt Indian firms

Even if this were to happen, it will take time.India will remain a good place to develop software


Outline

The software industry – an activity, not an industryHow did the underdog become a tiger?Prospects: Moving up the value chain

ImpactNow and futureEmployment?Intangibles: catalyst, exemplar and legitimizer


Impact

Direct Impacts Modest

Employment ~ 1 million35% of exportsIT + BPO ~ 7% of GDP by 2008Clearly such growth in the future is unsustainable but no reason for a crash in the near future

Indirect Impacts more important

LegitimizerBrand IndiaHelps other industries such as semiconductors, BPO, engineering services ..

CatalystChanges in corporate governanceCapital market listing requirements

ExemplarGet rich honestlyDo not require connections or privileges to succeedAre not inferior“software was virtually the first instance where wealth was created honestly and

legally, and more important, visibly so. Before this, wealth came either from breaking laws or at least bending them to one’s convenience, using existing

political and economic power. Hitherto commercial success had invited envy, cynicism and even outright hostility, and only rarely, admiration. While envy

and hostility are by no means gone, there is much more of admiration, and more importantly, a desire for imitation.”

Conclusion: From Underdogs to Tigers 2005


Origins of leading BPO firms: (Domestic SW origins highlighted)

Empl Start Year SW Origin

Genpact 26000 1997 Spin-off (GE)

IBM Daksh 18000 2000 Yes

Wipro 16000 2000 Yes (Acquired Spectramind, a startup)

WNS 10000 1996 Start up – Diaspora

Convergys 10000 ? MNC

HCL BPO 10000 2001 Yes

Intelnet 9500 2000 Startup – HDFC

Mphasis 8300 1999 Yes Spin-off (Citibank)

Aegis 8000 2004 Business House (Essar)

Sutherland 8000 1986 MNC

Hinduja TMT 7500 2001 Business House

ICIC One-Source 7300 2001

EYesL 7300 1999 Start up – Diaspora

Progeon (infosys) 7000 2002 Yes

24/7 7000 2000 Startup – Diaspora

TCS 5000 ? Yes

Efore 3200 1999 Startup-Diaspora

Vcustomer 3000 1999 Start up – Diaspora

Sitel India 3000 2000 MNC

Transworks 2235 1999 Business House (Birla)

GTL 1700 1999

Datamatics 1125 1991 Yes

Techbooks 1988

Efunds MNC

Data

qu

est

top

BP

O fi

rms,

2005

Sou

rce: T

he list o

f BPO

firm

s an

d th

eir size

is from

Data

qu

est m

ag

azin

e, th

e o

rigin

s are

base

d o

n in

form

atio

n a

vaila

ble

from

com

pan

y w

eb

sites


Thank You

Click here for slides

Documents

Transcript of Click here for slides