India and the IT industry

By Rahul Srivastava

India has been extremely successful at providing services to the world, from developing software to manning call centers. There have been many theories as to why, but there is no single answer.

For example, knowledge of the English language is often noted as a key Indian (and Irish and Israeli) advantage; the size of the workforce is another, but the success of Ireland and Israel, which employ only about 30,000 engineers, each belies this. The Philippines’ 20,000 software engineers speak English in the American style and work in an environment with better and cheaper telecommunications, roads, and electricity. Yet they are well behind India in the quality of the software they develop. Even in simple call-center work—where American English may be considered as an overriding advantage—the Philippines lags behind India.

Infrastructure is another negative for India relative to the other countries named. Also, China, keen to catch up with India in software, offers government support and superior infrastructure at a lower cost. It has stronger connections to technology and venture capital companies via the diaspora. There are more than twice as many engineers of mainland-Chinese origin in Silicon Valley than of Indian origin.

China produces a better quality of software engineer than India, if one goes by the patent record. Even so, innovative companies like Yahoo! and Adobe are shifting their highest-quality work from all over the world, including China, to India.

Indian government policy has, generally, been unhelpful. In the 1970s, when India began exporting software, policy was statist and protectionist. The state seemed to actively hinder private enterprise in software. It imposed high tariffs on the imports of hardware and software and discouraged foreign companies from operating in India.

Another key talent that is needed, entrepreneurship, perhaps we shall emerge from the quicksand of unsatisfactory explanations. Indian industry has created a class of determined and innovative entrepreneurs—even if they were not the most technically up-to-date in all fields. That is because after independence entrepreneurship had to be highly productive in order to offset the effects of hostile government policy, crumbling infrastructure, and expensive capital.

With no significant domestic market and no exposure to world markets, it was impossible for India to conceive of a new software product like the Microsoft operating system. Hence, Indian software began as a custom service—that is, software programs were written to specifications provided by their customers.

To the retail user reliant on mass-produced packaged software like Windows, this might seem like an expensive way to produce software. And it is. Despite this, the global custom-made software industry as of 2007 is worth about $400 billion a year, about twice the size as the packaged-software industry. The reason is that large companies in retail, banking, insurance, telecommunications, and others who view their offerings as proprietary regard the software that is behind these offerings as a competitive tool. For example, a large bank such as Citigroup that offers its customers online banking wants to make its service better than the competition. For this, it needs to customize the service in ways that distinguish it from others. This almost always requires writing software that is specially tailored to the service. It costs more, but it is willing to pay the price.

India did not invent the writing of custom software. The business was invented by companies like GE trying to capitalize on IBM’s decision in 1969 to separate the selling of hardware from software. IBM did this by specifying open standards for its hardware. An independent software provider could use these standards to develop software that would work on IBM’s machines.

After IBM made that fateful decision, the software world was never the same. Earlier software firms survived in an IBM-dominated world by providing simple services such as computer maintenance and time-sharing. After 1969, they could also provide software for operating the computer and for specific industry applications that IBM’s software engineers might not have written well. A large independent software vendor (ISV) industry quickly developed. By 1974, when India began exporting software, the ISV business was already well developed in the West, particularly in the United States.

At that time, the large mainframe manufacturers, such as IBM and Burroughs, were established in India. Burroughs had a joint venture with TCS, a division of India’s largest industrial group, Tata Industries. As S Ramadorai, the CEO of TCS told me, “Burroughs soon noticed that our engineers did an excellent job

of installing and maintaining Burrough’s products in India. So they asked us if we would send a few of our best engineers to the US to do the same for Burroughs’s clients in America.” Thus began the business of software exports. TCS sent programmers to the United States to install the systems of Burroughs. The industry’s term for the business was body-shopping. It was an odd term because the best minds, not the best bodies, of the country were being sent overseas.

The Indian software industry thus began by recruiting for western companies. It sought talent that could mimic western programmers using the same production techniques. The difference is that costs were lower than the West’s owing to India’s low labor costs and its large labor pool. This approach, termed labor arbitrage, had already succeeded in East Asia a decade earlier, though in manufacturing. It was to be replicated with even greater success by China—again in manufacturing—a few years later.

However, there were some differences between China’s mass-based manufacturing and Indian software exports. The first difference was that, for the first two decades, software exports from India were not exports of written code. Instead they were exports of people to clients’ sites in the United States and other places to write code.

Second, the technology exported to China was, for the first two decades, simpler than what was used in the West. The output that emerged from Chinese factories was of second-tier quality. By contrast, Indian programmers, from the very beginning, worked side by side with western programmers in the same work environment and used the same equipment.