a de facto rent subsidy in a technology park.
ā— Governments ease the bureaucratic process. This is done by giving the investing
company specialized attention, such as one-stop investment and business
registration processing, and simpliļ¬ed export and import procedures.
Some of these incentives are dependent on speciļ¬c geography, speciļ¬cally in technology
parks or in tax-free zones, while other incentives are not speciļ¬c to location.
Countries that rank in the top of āR&D subsidies and tax creditsā are, in order:
Israel, Singapore, Taiwan, Canada, and Ireland.13 Not surprisingly, none of these most
inviting nations are low-wage nations that compete largely on labor arbitrage.
Some examples of the range of government incentives appear in Exhibit 4.1.
These incentives should not determine your companyā™s investment choice, but rather
they help tip the scales after narrowing country choices to the top two or three
Costa Rica ā“ 8-year income tax exemption, followed by 4 more years at half of the standard
30% tax on proļ¬ts; duty-free imports; unfettered capital repatriation.
India ā“ Software proļ¬ts are not taxed until 2009, 10-year sales tax exemption, 5-year tax
exemption on imported capital goods, special concession for power generation.
Ireland ā“ Cash incentives of up to 45% of the investment. Standard 10% corporate tax rate.
Israel ā“ Cash grant of 24% of investment, or a 10-year income tax break.
Malaysia ā“ 10-year tax holiday on proļ¬ts and tax exemption on imported capital goods.
Mexico ā“ 30% of R&D expenses are tax free.
Philippines ā“ Software ļ¬rms reside or register in high-tech zones (which can be in Manila or
elsewhere). Once the ļ¬rm is classiļ¬ed as an export-oriented ļ¬rm it receives a 4-year tax
holiday and duty-free imports.
South Africa ā“ Foreign software ļ¬rms that conform to the āsmall and medium size enterprise
development programā receive cash grants on a sliding scale, for example the ļ¬rst R5 million
(about 500,000 USD) are matched up to 10%.
Exhibit 4.1 A sample of government beneļ¬ts and incentives for foreign technology companies
77 The offshore country menu
Case study Sport Systems Inc. shops creatively for an offshore provider
āFrom my previous experience with offshoring to India and to Vietnam I learned
that there were not huge differences between the providers in each country ā“ but,
rather, that each country was very different: how programmers solved problems
in each country was quite different.ā (Charles Angler, of Sports Systems Inc.
(SSI), on why he decided to test providers from three different countries.)
This is an actual case. At the request of SSI all names are disguised.
Once SSI decided to develop its new software product offshore, it came up with an
unusual plan. The company chose three offshore providers, one each from India,
Russia, and China, to compete for the project. SSI planned to contract with each
provider to complete an identical ļ¬xed-price pilot project. Comparing the results,
SSI would choose the best performing provider for the full project.
SSI is a 20-person American software product company specializing in a distinct
segment of the sports business. While the ļ¬rm has been successful, it competes
with many other ļ¬rms in this segment. Charles Angler joined SSI in 2003 as
Director of Systems Architecture. In his previous job he gained ļ¬rst-hand experi-
ence in offshoring. Some of his projects succeeded and others failed. Angler talked
about his lessons:
āWhen youā™re in the negotiations phase, and you talk to these offshore
companies, which you have never met, theyā™ll all say they can do it, theyā™ll
say we have all these wonderful references, hereā™s some sample code ā“ and
it all looks good ā¦ but you want to look at the substance ā¦ā
It was clear to SSI and to Angler that, given SSIā™s small size and limited funds, it
could not develop its new product inside the company or pay the high rates that an
American ļ¬rm would ask. Angler identiļ¬ed and began negotiating contracts with
three providers: Powercode Plus of India, TQ-Link from Russia, and Sun3 of China.
āThe key success factor to offshore work is to create strong process ā“ in particular,
strong speciļ¬cations using uniļ¬ed modeling language (UML)15 and make sure that the
offshore provider is experienced in iterating with UML,ā said Angler. Together with
one of his business analysts Angler carefully speciļ¬ed the project and divided it into
ļ¬ve releases. The pilot was made up of a subset of āuse casesā that span the full process
lifecycle. The pilot included a project plan, a graphical user interface, a UML-based
logical view, a design for integration with Web Services, and a ļ¬nal .NET prototype
(including test cases and results), with corresponding iteration plans throughout.
The SSI offshore contest was now ready to begin.
A problem immediately arose in contract negotiations. The Indian company,
Powercode Plus, became too fussy about how it would treat the software code in
78 The fundamentals
case of dispute between the parties. SSI, a small player in a ļ¬ercely competitive
segment, was very concerned about protecting its intellectual property, leading it to
abandon the negotiation and eliminate Powercode from contention.
The Chinese company, Sun3, was very aggressive from the start, completing the
project in 14 days. āThey were working all hours,ā said Angler. āI knew because we
required the offshore team to be on instant messager (IM) all the time ā“ and they
were almost always there, even with the enormous time zone differences.ā In fact,
Angler was concerned that they worked too hard and would not be able to sustain
this type of effort over a longer, bigger project.
Angler and others at SSI reviewed Sun3ā™s code and concluded that it demonstrated
reasonable skills in UML, although the offshore team was not familiar with some
best C# coding practices. Sun3 programmers had created some object classes that
were too broad and might lead to extensibility problems down the road. But, once SSI
explained the deļ¬ciencies, the Sun3 team quickly brought the source code in line
with standards. Sun3 completed the pilot for 2000 USD, using 150 person-hours,
based on a charge rate of 11 USD per hour. āThe interesting thing,ā said Angler, āis
that during the entire project duration, we never once had a voice conversation with
anyone at Sun3 ā“ everything was through e-mail and IM.ā
Meanwhile, the Russian company, TQ-Link, took 1 month to complete the project.
āWe were very impressed with their work,ā said Angler, who noted that TQ-Linkā™s
work came in solid the ļ¬rst time around. TQ-Link demonstrated deeper knowledge of
UML, produced a somewhat better user interface, and built a rigorous test suite. āAnd,
I had excellent communications with their Boston-based representative, Sergey, who
speaks excellent English.ā TQ-Link completed the project for the ļ¬xed price of 8000
USD, using 425 person-hours, based on a blended charge rate of 20 USD per hour.
āWe were quite happy with both companiesā concluded Angler. But, although
TQ-Link was superior on several dimensions, the differences between the two ļ¬rms
were small. This magniļ¬ed the cost difference, since the Sun3 rates were about half
as much as TQ-Link. SSI chose the Chinese ļ¬rm, Sun3. Angler added: āI suspect
that the actual hours, for Sun3, was probably closer to 250 hours than 150.ā
Replicating several small pilot projects is an exemplary approach for choosing both
countries and providers. It is worth considering for both ļ¬rst-time offshorers and
experienced offshore managers like Angler. SSIā™s careful speciļ¬cation reduced mis-
communication and focused the evaluation on the providerā™s true technical and
business capabilities. On the other hand, SSIā™s decision was not as clear-cut. It was
faced with two good choices. It seemed to choose cost over quality. Not all ļ¬rms
should make a similar choice.
79 The offshore country menu
Country sketches: the Big Three and eight more
We would like to give sketches of the 100 nations that export software. For practical
purposes, we chose eleven nations that represent something of a cross-sample. We
begin with the Big Three: India, China, and Russia; then cover the two other āIā coun-
tries brieļ¬‚y, Israel and Ireland. We then introduce six relatively unknown offshore des-
tinations: Latvia, Romania, Malta, Vietnam, Bangladesh, and Costa Rica.
The country sketches begin with the āBig Threeā offshore destinations: China, India,
and Russia. Much of the volume of software offshoring is going to these three large
nations. These nations are also big in the traditional sense, in that all three have large
populations. In addition, these large populations also have a large, well-educated labor
force in Science and Technology (S&T). In current industry parlance, we say that these
nations have āscaleā and a ādeep labor pool.ā All three have seen substantial government
investments in human capital (literacy, schools, and universities).
There are a number of other important common denominators among the Big Three.
Of course, all three are already active in offshore software work, and in all three wages are
low relative to industrialized nations. All three have opened up their economies consid-
erably since 1990. Yet, all three nations are deeply divided economically, with computer
and Internet penetrations at relatively low per-capita levels. All three have software
organizations that are very rapidly improving processes and software quality. All three
have seen most of the innovative software work take place inside the captive centers
of foreign technology companies, rather than their own home-grown ļ¬rms. All three
have had little commercial success with independent indigenous product R&D. All three
have poor protection of IP.
Broadly, India has become successful as a software factory, perfecting software pro-
duction and delivery systems. Russia has had some success in algorithm-oriented soft-
ware that requires invention and resourcefulness ā“ generally called software R&D.
With Chinaā™s software export industry so young, many have been wondering if it will be
factory- or algorithm-oriented. It is not yet clear where China will turn. Chinaā™s success
in manufacturing suggests that it will be successful in factory orientation. Chinese tem-
perament suggests that it will be more successful in algorithms. It is quite possible that
China will succeed in both.
We begin the country sketches with the Big Three nations.
Indiaā™s success in software has been so extraordinary that it has been the subject of
many books, and countless magazine and newspaper articles trying to make sense of
how this poverty-stricken country became a glitzy, dynamic, high-technology power-
house that makes European executives and Washington policymakers fearful.
80 The fundamentals
Indiaā™s remarkable success is driven by multiple factors. First, and most important, is
its vast human capital that is well educated and English speaking. The national network
of quality universities has been one of its gems. The elite Indian Institute of Technology,
of which there are several campuses, accepts 3500 of 178,000 applicants a year, a
selectivity rate of 2%. The scale of human capital is enormous: by some estimates, there
are more IT engineers in Bangalore than in Silicon Valley (150,000 versus 120,000).
Indiaā™s industry has successfully created a top layer of large, dynamic, multinational
ļ¬rms. Not just one or two successful large ļ¬rms, but several. This is a feat that the
Russians, so far, have failed at entirely, while the Chinese industry is still too young
to judge. The top Indian ļ¬rms (usually labeled Tier-1 ļ¬rms, introduced in Chapter 1) are
extraordinary success stories in their own right. By 2004 there were three Indian IT
firms with more than 20,000 employees and two more above 10,000. The top Indian ļ¬rms
have also been unusually proļ¬table, with gross margins of 45% and net margins of 30%,
all while growing at 32% between 1999 and 2004. The ļ¬rmsā™ success in IT services is
epitomized by capturing business with more than half of the largest US corporations
(See Figure 4.2).
India has also been successful at creating synergies between similar knowledge-based
sectors: it began with software services and used these competencies to move into
software R&D and ITES. The ITES market has grown from near zero in 1998 to
2.4 billion USD by 2003, and is forecasted to grow eight-fold within 5 years. Indian
ļ¬rms have also been quite successful in software R&D, estimated at 1.3 billion USD
in 2003, with 77 global ļ¬rms having established direct R&D subsidiaries in India.16
India has also grown a successful cluster of independent software R&D contracting
ļ¬rms (ālabs for hireā) that perform projects for foreign companies. The largest of these
is Wipro, with 6500 engineers in its software R&D division.
Number of US Fortune 500 firms
1990 1996 2000 2001 2002
Figure 4.2 Number of US Fortune 500 corporations with some offshore IT work conducted
Source: Authors based on NASSCOM data.
81 The offshore country menu
In summary, the story of offshoring is, in many ways, the story of success of the
The Chinese software industry was not on anyoneā™s radar screen as recently as 1999.
Like the modern Chinese cities that seem to sprout up almost overnight, Chinaā™s soft-
ware industry has emerged to become a global player in just 5 years. In China, more
than any other nation today, structural changes happen quickly.
Chinese human capital indicators are as impressive in quality and quantity as those of
India, with the exception of language skills (although there has been improvement in
recent years). Chinese universities produce roughly 300,000 engineering graduates per
year. While science and technology in universities has traditionally been too theoretical,
this too has been changing rapidly. Approximately 30ā“45 universities have launched new,
specialized schools in software. Such rapid adjustments show the ability of the govern-
ment to redirect resources on a massive scale. Chinaā™s human capital has been augmented
by a reverse brain drain. Approximately 160,000 Chinese have returned with foreign edu-
cation. Many are starting ļ¬rms or choosing to work for foreign multinational ļ¬rms. These
returnees are bringing with them considerable know-how in technology, managerial proj-
ect and process experience, experience in western business, and ļ¬‚uency in English.
The Chinese software industry is not as distinct as the Indian industry in two
respects. First, the Indian industry does relatively little work domestically, while the
Chinese industry does a great deal. Second, the Chinese industry is more closely tied
to computer hardware and other manufacturing industries. Chinaā™s software strengths
have speciļ¬cally been in embedded software at the interface between hardware and
software: in telecommunications equipment, data communications, and wireless. Yet
growth has taken place in all major software segments: in services, product R&D and
embedded software, as well as in the related ITES.
The volume of exports of software products and services for 2003 was somewhere
under 1 billion USD, representing roughly 10% of the Indian software powerhouse, but
expected to reach 30% of the Indian volume by 2008.
Most large technology ļ¬rms from the US, Japan, Europe, and India have software cen-
ters in China. Unlike India, in which the USA and Europe are the dominant investors
and clients, Japan is a key investor and client in China.17 The majority of large
American technology ļ¬rms, including Cisco, Intel, Microsoft, and Motorola, do some
R&D work in China, including some innovative R&D. In total, R&D centers have
risen from 150 in 2002 to 400 by 2004.18 While Chinese software units have not attained
the large number of world-class quality marks (Capability Maturity Model (CMM))
that Indian ļ¬rms have attained, this is less signiļ¬cant for R&D activities.
Of the 2000 indigenous Chinese IT ļ¬rms, only a small number export software
services or products. And none of Chinaā™s pure software ļ¬rms have attained the size of
82 The fundamentals
the major Indian ļ¬rms. Noteworthy in size is Huawei, at the hardwareā“software inter-
face. This is perhaps the most interesting Chinese player, as a major competitor to Cisco,
with 2002 revenues of 2.7 billion USD (with almost one-third derived from exports).
Most exports are to developing/emerging markets rather than the rich markets of Europe
and America. Huawei has built a network of R&D centers outside China in India, Texas,
California, Sweden, and Hong Kong.19
The Chinese software industry beneļ¬ts from four complimentary growth drivers20
which will all interplay in its offshore software industry in the coming years. The ļ¬rst
driver is government-led development. Government support is quite strong: through
procurement policies and through its inļ¬‚uence at the national and local level. In China
this makes an impact quickly. The second driver is to follow the India model by pro-
viding IT services to foreign ļ¬rms. The third is to continue to attract Foreign Direct
Investment driven by the desire for market access by global ļ¬rms.21 The fourth and last
driver is to continue to be nourished by brain circulation ā“ the return of thousands of
Chinese bringing with them critical know-how.
The Russian software sector is by far the smallest of the Big Three nations. After the
break-up of the Soviet Union Russia was viewed as having enormous potential in soft-
ware but has not lived up to those lofty expectations.
Russiaā™s strength is in its large workforce educated in science, mathematics, and
engineering, including many with advanced degrees. Figure 4.3 points to the strength
in Russia in the proportion of advanced degrees versus India. This strength should con-
tinue to manifest itself in software R&D. While the educational pipeline produces sig-
niļ¬cant numbers as indicated in Table 4.2, the IT industry has absorbed little of the
output. In 2004 there were only about 70,000 workers in the IT industry as a whole,
with only an estimated 16,000 employees in the software export sector.22 The educa-
tional system has not been re-engineering itself as quickly as the Indian and Chinese
competitors. For example, as of 2002 no higher-education institution used the com-
plete reference models of ACM or IEEE for computer science or Management
Russian software exports are the smallest of the Big Three at roughly 350 million
USD in 2004. Of this amount, about 50 million USD came from captive software R&D
(i.e. owned by foreign ļ¬rms) and roughly the same amount in software products.
Indigenous software product ļ¬rms have not been inļ¬‚uential, with the exception of
Kaspersky Labs, a global provider of anti-virus software.
Russia has a respectable presence of American R&D centers including Intel,
Motorola, and Sun, as well as some European ļ¬rms, such as Nokia. Other foreign ļ¬rms
perform contract R&D. But, in total, this presence is signiļ¬cantly smaller than multi-
national corporationsā™ presence in India and China.
83 The offshore country menu
BTech MTech PhD MBA Other Other
Figure 4.3 Educational background of employees in Russian and Indian ļ¬rms.
Source: Bardhan and Kroll.24
Table 4.2 Russian educational systemā™s annual supply of IT labor
IT engineering graduates 42,000 46,000
Math and Physics majors 22,000 22,000
Non-IT engineering graduates capable of entering IT workforce 69,000 76,000
Graduates from other disciplines capable of entering IT workforce 71,000 81,000
Total fresh IT labor supply 203,000 226,000
Columns do not total due to rounding.
Nevertheless, the Russian software export industry has been growing steadily with
the consequent growth pains familiar in China and India: tight labor markets for the
best talent, increasing wages in the major cities, and high rents in desirable locations.
Russian ļ¬rms have been quick to emulate the Indian model by embracing international
quality standards such as ISO and CMM. Motorola Russia attained CMM Level 5 in
2001. One of Russiaā™s largest independent IT services ļ¬rms, Luxoft, attained CMM
Level 5 in 2003.
The internal domestic marketplace has become healthier due to the economic boom
of the early 2000s, and more sophisticated. This is mixed blessing for offshoring. As is
typical at this stage of industry growth in emerging nations, domestic customers are more
proļ¬table than foreign customers, diverting companiesā™ attention to domestic markets.
One of the most important limitations to industry growth is the relatively small size
of Russian ļ¬rms. Only a handful of ļ¬rms are larger than 200 employees. Individual
firms cannot ā“ or do not want to ā“ grow for a number of reasons: lack of capital (capital
markets are immature) and a desire to stay off the radar screen of the erratic and punitive
Russian tax collectors. There is an equally strong desire to stay off the radar screen of the
endemic organized crime and burglary rings. Finally, there is a Russian cultural pro-
pensity to work in intimate family-like organizations. The result is that software ļ¬rms
84 The fundamentals
choose to operate in a constellation of alliances, sub-contractors, and consortia that
allow them the ability to accept new work without growing their core workforce.
The other two āIās
The three software stars of the 1990s were the āthree ā˜Iā™s.ā India was covered earlier in
this chapter. We now provide brief sketches of the other two āIās: Israel and Ireland,
with the monikers of Silicon Wadi and the Celtic Tiger, respectively.
The two countries have several characteristics in common. Both nations are small
with strong educational systems. Both became successful, in part, in software products,
which is unusual for offshore nations. Both have seen their starring role in software
dull somewhat after the technology boom ended. Both attracted signiļ¬cant investment
from major global technology ļ¬rms (although into different segments as we discuss).
Both have relatively high wages that make them less attractive to offshoring of IT ser-
vices with its emphasis on wage differentials. Both have been hurt by the ļ¬erce global
competition as companies look to India and China. Both countries did not grow the
software ācontract R&Dā segment that India has been so successful at, preferring to
conduct software R&D on their own, or within foreign subsidiaries.
During the technology boom years of the late 1990s, Israelā™s software strengths were
in synch with global interest. With a highly educated labor pool and tuned-in technology
entrepreneurs, the country was able to grow many innovative software product ļ¬rms,
of which the most well known is Check Point, the information security ļ¬rm. Israel
became the number one foreign destination for American venture capital and the home
of more than 100 technology ļ¬rms listed on the US NASDAQ stock market.