OUT-OF-THE-WAY INFORMATION TECHNOLOGY

I continued my pursuit of information technology, and visited some more places unknown for IT; this series was published in Business Standard in June, July and August 2002. 

GETTING A FEEL OF SOFTWARE

I    IT entrepreneurs in Calcutta

Calcutta may seem an unlikely place to look for either information technology or entrepreneurs; but Anjan Ghosh is both. He comes from a village. His father was a trader, his mother a headmistress. This B Sc from Jadavpur got a job in a project: the World Bank had funded National Informatics Centre to computerize the state government. Anjan got a very cold welcome in Writers’ Building. The project room took ages to be handed over; in the meanwhile, the babus abused him for wanting to destroy their jobs. Then one day he went to a wedding. There he met two NRIs who had a distance teaching business in the US; they wanted someone in Calcutta to produce interactive CD-ROMs for their courses. Anjan took them around the city. After looking at many firms, they asked Anjan if he would do it himself.

Anjan did not know the ABC of business, let alone of interactive CD-ROMs; and his patrons did not give him much money. But while studying for the air force examination he had learnt that the keys to success were the three As - Attitude, Aptitude and Analytical skills. So he went to the local NIIT classes and got together about 20 people, offering them a share of his success. He did not have money for a swank office, and operated from a flat, worrying every day that the Calcutta Municipal Corporation would raid him and close down his business.

But he survived; he succeeded so well that the NRIs sold his outfit to DigitalThink, which sells training software in computer networks and telecommunications. Then two years ago, DigitalThink bought a training company with offices in Hyderabad and Bangalore; Anjan was sent to take it over. It was headed by an IIM graduate, who had stacked it with other IIM graduates. Anjan sacked most. But he found the programmers in the new outfit much better; they even wrote their own tools. So he brought some of them to teach his Calcutta staff.

I asked Anjan why DigitalThink chose him instead of the IIM alumnus to head the combined outfit. He said it was because he trained leaders. To train them you had to choose people better than yourself and give them space – but only limited time. They had to be given a deadline by which they had to prove themselves. Leadership was not learnt in IIMs; Anjan’s best deputy was a B Com.

Anjan has reduced training into an assembly line operation. Choose a subject, and throw together some subject matter experts and instruction designers. Make them preparee audio and visual content; send it to content integrators. Then send the courses prepared to delivery solutions, which goes through all the steps and makes sure the course delivers the training.

If Anjan knows the secret of computerized training, Brigadier Suresh Menon knows all there is to know about call centers – he runs the only call center in Calcutta for BNKe Solutions. Brigadier Menon calls his call center an infantryman’s dream, for its core is assertiveness training. The call center recruits girls aged 21-24 – the age after they have graduated and before they marry, when they want to put aside a nest egg. When they are recruited, their parents are invited to the call center to reassure them that their daughters will be looked after. It is a 24-hour operation; the girls are transported from home and back. The first two batches were trained by a Scotsman; then BNKe worked out its own training course. And the training is not all about acquiring an American accent. The girls must certainly learn to speak clearly. But the main thing is reassuring the customers – listening to them, letting them let off steam, and leading them to a solution in as short a time as possible – for time is the scarce resource, and productivity matters. All calls are monitored; if a girl makes a mistake, she is taken off the desk and to another room where she is taken through her mistake and shown how to do it right next time. If there are tears, Brigadier Menon is always around to take care of the girls.

Bikram Dasgupta is an entrepreneur on quite a different scale. He used to work in PCL in Delhi. The West Bengal government was trying to persuade Bengalis to come and start businesses in Bengal; N Bhattacharya, the then managing director of West Bengal Electronics Development Corporation, offered Dasgupta a floor in one of the buildings WEBEL had built. Dasgupta asked for a building; he built Infinity Building in the Webel industrial estate for Rs 100 crore. There – and now in nine locations – he gives engineers structured learning to make them ready for the market. In the last year the students would do projects; on their basis Dasgupta erected a software export business. The model worked beautifully till 2000. Then the IT boom collapsed, and Globsyn Technologies got into trouble. I told Dasgupta he looked relaxed; was he out of trouble? He said, no; but he had worked out how he was going to get out of trouble. His next project is Globsyn Business School to turn Bengalis into first-generation entrepreneurs.

Calcutta IT is full of Bengalis; it could hardly be otherwise. But amongst them is Pierre Marlard, the local CEO of Sema. Sema is one of Europe’s foremost IT companies; it configures customized IT systems for big customers. Its software runs French nuclear power plants, regulates power exchange across Western Europe, underlies the British railway network. It is likely that your cellular phone bill was generated by Sema software.

Sema was bought by Schlumberger two years ago. Schlumberger has been in oil well measuring equipment since 1927; it places instruments down oil wells, deduces from them how to get most oil out of them at lowest cost, and sells the information to oil producing companies. Sema provides it with backend solutions.

How did Sema end up in Calcutta? It had heard of horror stories about Bangalore. GemPlus went to Bangalore, and lost 40 per cent of its employees in attrition in one year; it got so disgusted that it went and set itself up in Dubai. Foreigners flock to Bangalore to recruit programmers; and Sema, which is big in specialized system software but unknown in the world outside, would be at a disadvantage in retaining them. So it went to Calcutta. Turnover has certainly been lower in Calcutta.

Will Calcutta become the capital of IT? If Swarup Roy of WEBEL has his way, it will; he is certainly working hard to make the place hospitable. And IT is the ideal industry for Calcutta. But IT is a mental game. To succeed in it, Calcutta needs to attract people from all over India. It was India’s most cosmopolitan city just 30 years ago. Under the CPM, it has become a boring bourgeois city. It should open itself to some diversity.

II    IT entrepreneurs in Poona

My last port of call was Poona, where I have influential friends – Naushad Forbes is the chairman of the Poona chapter of CII, and his secretary Shireen can open all doors.

Let me start with the bank of the Mutha, where there is a cremation ground for Gosains. This rather obscure Brahmin caste cannot be numerous, but the scattered cenotaphs suggest that it must have been cremating its dead on this shaded tract for some centuries. Between the Gosain ancestral ground and the main road I found a building that had seen better days, next to an engineering workshop. There, just above a printer of biri wrappers and calendars, I found Mr K S Natrajan of Mascon Computer Services (no relation of Mascon Global). He is the only IT entrepreneur I have met who admitted having had a rough ride in the past two years; his turnover has fallen by 70 per cent. He is obviously a veteran, with his white beard: he started in 1981 by compiling databases for power and telephone utilities. Then in 1995, an NRI came to him with a big job from a US client and asked for help; that is how Natrajan entered exports. Soon he had to set up separate offices for domestic and foreign business – not only because the latter grew too big. The quality standards required by the two markets were very different: the Americans wanted perfect work; the telephone company managers did not care if mistakes crept into directories. In 2000, his foreign business turned down; but the next year promises to be much better. Meanwhile, his biggest need is for venture capital. The bank which lends to him is not into risk-bearing, the capacity of his friends is limited, and no venture capitalist would touch IT just now. A Norwegian client had offered him $6.5 million, but he did not sell out because he thought the client might retrench staff; he did not want his people to lose jobs. Today that Norwegian outfit has itself closed down.

Another entrepreneur who had a tough time is Vivek Mannige; but his story has a happier ending. Unlike most of his compeers who talk like machine guns, Mannige thinks before he speaks. A 1974 graduate of Kanpur IIT, he worked in Telco for 14 years, and then set himself up as an industrial consultant. From the start he had set his eyes on creating products. But making products is capital-intensive. It requires computers and operating systems, on whose imports there were insuperable restrictions then. Products have a longer gestation lag, during which the team has to be fed. And selling products required a presence in the US market, an expensive proposition for a small company. CanBank would not give money except against confirmed orders, and there was no point in designing a product if it was going to be sold off even before it was made. Mannige made one product which he had to shelve because he did not have the money to support it; after the bitter experience he decided to make an expensive product which would have fewer users and hence would cost less to market. He made a consumer lending product for banks. It was adapted to various markets, and sold in ten countries. But it could not be sold in the US, where banks prefer a local vendor. He also filed for patent for another tool (a tool is a programme which writes better – more reliable, faster – software than a human being). Those products and related work gave Mannige’s company a reputation, and he has just sold it off. Now he has set up another company, AccelTree, which is better capitalized and can concentrate on making tools. Mannige was the only respondent who admitted to having used non-engineers. His best programmer was a first-year B Com student; unfortunately he wants to go and do his C A.

Mannige also had something to say about Indian and foreign customers. Products are sold in the US with a 30-day money back guarantee; they cannot in India because sales tax authorities make it difficult to get a refund. American customers buy a product and use it or return it; Indian customers complain about price, need more service, more demonstrations, and often want to return the product after 30 days. NIC asked for four demonstrations for a product, and then did not take it. Indian customers also buy foreign products more readily than Indian ones. But Mannige has better opinion of the Indian government. He sold his company to a foreign buyer; FIPB took only three weeks to approve, and Reserve Bank took only two days to e-mail queries after receiving the application.

Another IT entrepreneur who took the hard way was Anand Deshpande of Persistent Systems; but he did not enter the product market. Instead, he specialized in writing software for software producers such as Cisco and Agilent. They defined standards that were higher than in the services market; but once the standards were met, the customers did not need to be courted and buttered up. So Persistent had not used a single H1B visa in 12 years, and had done without an office in the US till very recently.

The market Persistent is in would normally be difficult for a small company in Poona to break into. But in 1992, when he was very hard up, Deshpande got an order worth $25,000 from Microsoft on onerous terms; it stipulated that if he failed to do the job in time, he would have to pay for the replacement. It was lucky he did finish before time, for otherwise he was prepared to go bankrupt.

With the kind of demanding work Persistent does, it would need very good programmers. He takes almost a fifth of the computer engineering graduates of Bombay IIT, and also recruits from Indian Institute of Science; he has seven Ph Ds in the company, including himself. The young graduates are bright; he puts them through a three-week training, and straightaway puts them on the job. Many go abroad for further study after a while; he encourages them to do so, gives them recommendations, and often helps them find a place. He publishes papers in academic journals and encourages his staff to do so.

One of his employees is his father, whose most admirable initiative is the Software Exporters’ Association. Its 70 members follow the law, and do not give bribes. If they face a problem with, for instance, a corrupt government employee, they take the matter to the top. Till now they have not encountered an insoluble problem. I wish CII and FICCI would follow in Poona Software Exporters’ Association; if they did, political and bureaucratic corruption would soon become a thing of the past.

III    A tremor or a portent?

Indian software sales, 1993-94 to 2001-02 ($ million)
	Domestic  sales			Exports				Total	Share of
	Software	ITES	Total	Software	ITES	R&D	Total	sales	exports
						services			%
1993-94	230		230				330	560	58.9
1994-95	242		242				485	727	66.7
1995-96	419		419				734	1153	63.7
1996-97	648		648				1085	1733	62.6
1997-98	1031		1031				1750	2781	62.9
1998-99	1224	21	1246				2650	3896	68.0
1999-00	1330	39	1370	3440	560		4000	5370	74.5
2000-01	1737	70	1807	4750	900	550	6200	8007	77.4
2001-02	1923	147	2070	5780	1475	575	7780	9850	79.0
Annual rates of change (%)
1994-95	5.3		5.3				47.0	29.8
1995-96	73.1		73.1				51.3	58.6
1996-97	54.7		54.7				47.8	50.3
1997-98	59.0		59.0				61.3	60.4
1998-99	18.8		20.8				51.4	40.1
1999-00	8.7	83.6	9.9				50.9	37.8
2000-01	30.6	78.7	31.9	38.1	60.8		55.0	49.1
2001-02	10.7	110.0	14.6	21.7	63.9	4.5	25.5	23.0
1994-5/2001-2	30.4		31.6				48.4	43.1
Source: Nasscom.

When I first went to Germany, I used to think that Germans were the world’s cleverest people; why, even their children could speak German. I was therefore quite shocked when I visited Germany last year and a German said to me, “You Indians are very intelligent, aren’t you?” It took me a moment to realize that he was talking of our information technology industry. IT is justly an industry of our pride. In a country short of achievement, it has achieved incredible growth; going out of a country afraid of the outside world, it has captured world markets. And it has all been done by ordinary Indians, most of them born and educated in this country.

Those of us who take a sanguine view of this industry should take seriously its sales figures for 2001-02, released by Nasscom and given in the accompanying table. For its sales growth has fallen drastically from 49 per cent in 2000-01 to 23 per cent last year. The fall is spread through both the domestic and the export markets; domestic sales growth fell from 32 to 15 per cent, export growth from 55 to 26 per cent. Normally, if one market weakens, an industry sells in another; but obviously the domestic market is not making use of the slack in the industry to accelerate purchases.

In both the domestic and the export markets, growth has been pushed up by the rapid expansion of IT-enabled services. These are a very different kettle of fish from software services; young women making cold calls to American housewives trying to sell them digital cameras are miles apart from programmers writing code to fly a plane or monitor an oil well. If we take out IT-enabled services, sales growth of software fell to 11 per cent in the domestic market and 22 per cent in exports – still high rates, but a far cry from the rates of 50 per cent and more prevalent just a couple of years earlier.

Meanwhile, Nasscom and McKinsey, in their updated forecast, have brought down expected growth only slightly. Is the fall in growth in the past year just an aberration, or does it denote the end of the golden age, and will it be all downhill from this point?

It should be recalled that the figures in the table are in terms of millions of dollars, and are affected by changes in volume as well as in price. The hourly rates charged by Indian software exporters have fallen in the past year. Insofar as they have, the volume of sales would have fallen less than the value. If the rates fell 20-25 per cent, the volume may not have fallen at all; if they fell more, the volume may have gone up. Nasscom’s estimate of the fall in rates in the first quarter of 2002 was 9 per cent; in which case volumes must have fallen by 10 per cent at least.

Which is the case is relevant to a difference of opinion at the outset of the current slowdown. When the American IT industry melted, some people thought that the Indian industry would melt with it. Others thought that since the Indian industry’s costs were lower, it would take away business from the US industry and might actually grow. What really happened?

I have been talking to a large number of Indian IT firms – on some of which I have written in previous columns. I have written mostly about the successes; I have come across some failures, or some firms in trouble. Interviews are not the best way of finding out what is happening to an industry. For one thing, one cannot meet firms that are dead already. For another, human beings find it easier to talk about their successes than failures. So one is likely to get a rosier view than reality would justify.

The firms that survive by and large talk of a bad year 2001; their downturn started some months later than in the US because of contracts in hand. They talk of recovery in 2002 – not just the prospect of recovery but actual orders beginning to flow again. They admit to having cut rates, but not by all that much.

But most of those I talked to had one niche or other. Less specialized firms have been much worse affected. In general, the business of providing programmers, whether directly or in the form of jobwork ODC (other direct cost) contracts, has been hit. I do not know how much they cut their rates; but whether it was 5 per cent or 50, it was not enough to save their bacon.

I have just been to the US west coast; there, stories of Indian programmers having lost their jobs and taken pay cuts are legion. But it is clear that very few of them have returned. Some no doubt have – especially those with H1B visas who could not find a job quickly enough to retain them. But the numbers are probably more like 5,000-10,000, not 50,000-100,000. Remember that there are some 200,000 Indian programmers in the US, against perhaps 300,000 in India; if a sizeable number came back, they would make waves here. Indian firms have been recruiting in the US; but most of the recruits are either specialists, or management cadres, or people with contacts in markets. They are not junior programmers.

This suggests to me that whilst there has been a shakeup in the US software industry, and many firms have shut down or downsized, total employment has hardly fallen. The US industry has itself cut rates – not always by charging less for the same job, but by taking on bigger jobs for the same price.

In general, they have been going to their clients and saying, “Why do you maintain these huge IT departments? We will run them for you.” They have been taking on functions of their clients which are information-intensive. This business process outsourcing is seen by many Indian firms as a great opportunity. But it has happened already; and most of the business has been captured by local American firms.

The movement has not reached exhaustion; there is still more business to come. But the government has put Indian firms at a great handicap in getting it; for first its passivity in the face of Gujarat riots and then its belligerence against Pakistan have led India to be perceived as an unsafe country, and few American or European businesses would now contemplate outsourcing business processes to firms in India. This by itself has meant a huge loss of business. If it were 10 per cent of current software exports, it could amount to Rs 50 billion. Add to it the cost of confrontation.

The confrontation will not go on forever; nor can daggers always be drawn in Gujarat. But the long-term growth of the software industry in this country will continue to depend on who is in power and what use he makes of it.

IV   IV    What is this software?

Composition of software exports, 2000-01 and 2001-02 ($ million)
				Exports		Domestic sales		Total sales
				2000-01	2001-02	2000-01	2001-02	2000-01	2001-02
Software and services				4750	5780	1737	1923	6487	7703
Legacy applications				1700	2100			1700	2100
Custom application development				1950	2350	352	358	2302	2708
Packaged software integration				300	350	426	442	726	792
E-business solutions				550	600			550	600
Wireless integration				75	100			75	100
System integration				75	110			75	110
Network infrastructure management				50	65			50	65
Consulting				50	55			50	55
Turnkey projects						362	358	362	358
Captive development						438	586	438	586
IT enabled services				900	1475	70	147	970	1622
Customer interaction centres				185	350			185	350
Business process outsourcing				295	600			295	600
GIS and engineering services				350	450			350	450
Other				70	75			70	75
								0	0
R&D services				550	575			550	575
Training						511	379	511	379
Total				6200	7780	2318	2449	8518	10229

India has become a mini-superpower in software; what is this software? Let me begin with Nasscom’s breakdown of software sales. Nasscom is admirably prompt; it has already put out estimates for 2001-02. According to it, software sales in the last year just crossed $10 billion: $7.78 billion in exports and $2.25 million in domestic sales (I have converted the domestic sales from Rupees to dollars). I have some doubts about these figures. For one thing, as the accompanying table shows, the increases in exports are in suspiciously round figures - $400 million in legacy exports, $25 million in wireless integration and so on. For another, Nasscom has included what it calls captive development – software produced in-house by companies that use it themselves. This departs from international practice; self-consumed software is not included in global sales. There are good reasons for its exclusion. It is not sold; hence its value is impossible to estimate. Nasscom has estimated it from manpower inputs; but that is still very approximate. And with self-consumed output, it is impossible to distinguish between software services – such as repairs and maintenance – and genuinely new software. If captive development is excluded, the growth of the domestic market in the last year comes down from 7 per cent to 3 per cent.

Thus the domestic market has hardly grown – whereas exports rose by 25 per cent. Ever since the start of the software boom, exports have risen faster than domestic sales; in the last year also the share of domestic sales continued to fall. This despite the fanfare surrounding the IT Action Plan of 1999 and the actions, supposedly taken, to increase the domestic use of IT, and despite the international slump.

There is some obfuscation of exports in Nasscom’s figures as well. Real software exports rose only 22 per cent; the growth was pushed up by including IT-enabled services, whose exports rose by 64 per cent. ITES use none of the skills required for software; but if they earn foreign exchange like software, there is no reason to scoff at them.

Over three-quarters of India’s software exports come from two sources: legacy applications and custom application development. Legacy applications arise from the use of older software. Big foreign companies have been using computers since the 1960s. The programming languages of that time, such as COBOL and Fortran, have been superseded by new ones like C# and Java. But where crucial data have been stored for decades on old systems, it is cumbersome and risky to scrap them and replace them with modern systems. So the big companies engage IT firms to patch together programmes and adapt their old – “legacy” – systems to cope with growing needs. Indian firms have traditionally specialized in this work, perhaps because Indian software engineers had only legacy computers to work with till the end of the 1980s.

The other Indian speciality – custom application development – is closely related; it involves working out software solutions for individual clients. These too are big, international companies; only they can afford custom-made solutions. Indian software firms have been particularly successful in serving banks and financial institutions. They have been doing similar work for local companies as well; the skills spill over into the domestic market. Often, new package programmes become available which offer better business solutions; then they are built into existing “legacy” systems. This type of packaged software integration is also done for both foreign and Indian firms.

These specialities reflect the biggest source of difference between Indian firms and their competitors abroad. The latter depend on and emerge by solving local problems – by designing controls for missiles in Israel, or computerizing pharmacies in England, or helping multi-location retail stores with their logistics in the US. Very little of this type of work is available in India, and Indian firms do little of it for local firms. The domestic market provides a proving ground for IT firms in other countries; it largely fails to do so in India. It is an achievement of Indian firms that they have managed to build up their business without a domestic market.

At the same time, the lack of a domestic market has been a considerable handicap for them. For instance, many firms here and abroad have tried their hand at enterprise resource planning (ERP). But the world leaders are American, German, British – but not Indian. That is because Indian firms do not provide good proving grounds for ERP software. The reasons are twofold. For one thing, the processes of few Indian firms are world-class; generally, their technology and management lag behind world practice. For another, implementation of IT solutions is much faster in client firms abroad than in India. They can retrench workers more easily, and transfer them from one function to another; the labour force in Indian firms is too inflexible.

Also, foreign companies are more tightly managed; once a decision is taken to change processes or the organization, they implement it quickly. Typically in India, a company has an IT department which buys IT; the implementation of what is bought is cumbersome, and involves slow and often infructuous negotiations between the IT department and the operational departments. The reason for the underdevelopment of the IT market in India is that management in India is underdeveloped.

This handicap becomes particularly important when it comes to a new generation of work based on internet or intranets. Before the arrival of internet, a company had to set up its own communication network and connect its computers across countries and continents if it wanted to set up an integrated management; it was expensive, and few companies did it. But VSAT and internet have  greatly reduced the costs of connecting up remote factories and offices; so there has been a boom in communication-related work – e-business solutions, wireless integration, system integration and network infrastructure management. 

Internet is little understood in India; we think that progress is measured by how many people play games or send e-mails. The real use of internet is in business communication. And for that use, fast and reliable telecommunication is the key. If this were understood, the government would cease to be concerned with internet users, and would instead provide a free broadband backbone to the entire country. 

It is because of the weaknesses in domestic infrastructure – the backward industrial and management structures, the underdevelopment of broadband – that the software industry has concentrated on the export market; there too it is handicapped because it cannot try out new solutions and technologies in the domestic market. After growing at 50 per cent a year for years, the growth of exports has halved in the past year. This is not just due to the global slowdown; it also reflects the industry’s disadvantage at home. Exports need a domestic market – not a large one, but a challenging and demanding one.

V    India’s market shares

Global IT services market and India's share 2001
	Global market		India's exports	India's share
	$ bn	%	$ bn	%
Professional services	142.9	32.5	5.3	3.7
IT consulting	21.3	4.8	0.1	0.3
Systems integration	81.1	18.4	0.1	0.1
Custom applications	19.3	4.4	4.5	23.1
Network consulting & integration	21.2	4.8	0.7	3.3
Product services	117.9	26.8	0.4	0.3
IT training and education	25.5	5.8
Hardware support & installation	44.4	10.1
Packaged software support & installation	48	10.9	0.4	0.7
Outsourcing services	179.2	40.7	0.1	0.0
Processing services	78.4	17.8
IS outsourcing	64	14.5
Application outsourcing	13.4	3.0
Network infrastructure management	23.4	5.3	0.1	0.3
Total	440	100.0	5.7	1.3
Source: Nasscom.

Last week I reviewed the specialization of the Indian software industry – what it did best and where it earned most. This time I want to look at India’s share of the world market. Total world sales of software in 2000 were estimated at $ 440 billion; India’s exports in 2000-01 were $ 6.2 billion – 1.4 per cent of the global market. If the world market had not grown at all in the following year, India’s share last year would have been 1.8 per cent – a bit higher but still modest.

Of the $ 440 billion global sales, $ 219 billion – almost a half – were in the North America, where India sold $ 3.9 billion’s worth in 2000-01; it thus had a 1.8 per cent share there. Sales in Western Europe were $ 127 billion; there India had a share of 1.2 per cent. In the Japanese market worth $ 53 billion, India had a mere 0.4 per cent share. The rest of the world did not matter; its market share was less than 10 per cent. Of India’s exports of Rs 283.5 billion in 2000-01, the US took Rs 173.3 billion, the UK Rs 33.6 billion, Singapore Rs 5.4 billion; 80 per cent of the software went to Anglophone and Commonwealth countries. But India has a respectable presence in all markets except the Japanese. One indicator is its subsidiaries and offices abroad. In 2000, 212 firms had set up 509 overseas offices – 266 in North America, 122 in Europe, 59 in Asia, 25 in Africa and 12 in Latin America.

India really caters to the English-speaking part of the world – the US, Canada, Britain, Singapore, Australia, New Zealand, Hong Kong. Its sales to Japan and Latin America are modest. The Japanese themselves have produced almost all the software they need – for their robots, consumer durables, animations etc – and are now recruiting large numbers of Chinese engineers for both offshore and onshore work. Indian firms are just getting interested in the Japanese market.

However, this impression of uniformly low market shares disappears when one looks at different kinds of software. Within the US, Commonwealth and Anglophone markets, Indian software exports are strongly concentrated in custom applications. As the Table shows, IT services are divided into three groups: product services, which are related to standard products, outsourcing services, where the supplier takes over one or other function from the client, and professional services – a residual category of services where an outside firm is called to provide a one-off service. India has virtually no presence in the first two. Since its hardware industry is insignificant and devoid of innovation, India is absent in hardware-related software. Indian firms have just begun to sell products, mainly financial ones, and to generate  revenue from servicing them; but the figures are very small. Thus in the entire product-related services, accounting for 27 per cent of the market, Indian sales are negligible.

Outsourcing of IT-intensive functions is the other big market, accounting for 41 per cent of global sales, in which India has virtually no presence. The only part of this market in which Indian firms have a small role is in the servicing and maintenance of web sites; but the income from it is insignificant.

The only segment in which Indian firms count is professional services – services sold, usually on a one-off basis, to help big clients manage and modify their IT systems. Here too, their presence is negligible in consulting and system integration, whether related to networks or not; they are still not called by big clients to tell them what to do, or to revamp their entire IT systems. They are more generally called for incremental services – improving or augmenting existing systems in a predefined direction. 

Aside from some banks and financial institutions, big clients do not call in Indian firms for jobs that are central to their business. They do not ask Indian firms for advice on how to change their IT systems or to adapt them better to their businesses. They do not allow Indian firms to handle core functions which must be kept running for their reputation or existence. This is the major constraint on the expansion of Indian firms’ market.

Indian firms are aware of this barrier; a number of those I interviewed mentioned it. Some attributed it to a cultural bias amongst western clients, others to doubts about India as a source of indispensable services. Indian firms have been aware of the huge market for outsourcing, and some of them have actively tried to enter it. But the results have been modest till now. Those clients that have outsourced processes to India insist on drastic steps to ensure security. They require that the supplying firm isolate the section serving them from their other business. They ask for multiple channels of communication, so that the failure of one will not stop information flows. And they worry about uncontrollable events such as wars, hurricanes and riots. The Gujarat riots and the tension between India and Pakistan have done much to put off outsourcing clients. In general, Indian firms are finding it hard to break into the outsourcing markets for reasons that are not in their control, but have much to do with the political complexion and foreign policy stance of India.

The prescriptions for the industry point in two different directions. Many people, especially those with little knowledge of the industry, argue that the industry must “move up the value chain”. What they mean is that within professional services, the industry must do more comprehensive jobs: advise clients on how to manage, set up and modify computerized management systems. They believe that it is firms abroad specializing in such a comprehensive service that make big money, while the Indian firms they subcontract only get labour costs plus a margin.

I think this view is mistaken. Cheap Indian programmers are there for anybody to use; if Indian firms do not provide services based on them, foreign firms will come to India and do it for themselves. Many do so already; with the coming of broadband, they will not even have to come to India to use Indian programmers. It would be obviously advantageous for Indian firms to break into the consulting and professional services market. But it is dominated by a small number of powerful firms, many of which give other services as well to their clients. Breaking into this market should be tried – and will be by the leading Indian firms – but is very difficult.

Far easier would be entry into the outsourced or IT-enabled services market. It is a huge market – total sales in 2000 were $179 billion. Its skill requirements are low – much lower than those of the software industry. Its employment potential is vast – it can employ hundreds of thousands of youngsters with modest education. And it can keep up the high growth momentum of the IT industry.

VI    Software manpower

IT manpower estimates (thousands)
			1999-00	2001-02	2004-05	2004-05
			Actual	Estimate	Optimistic	Minimum
Employment of IT professionals
Software exports			110	170	1060	641
Software domestic sales			17	22	49	38
In user organizations			115	224	300	250
			242	416	1409	929
New entrants
IT graduates				57	79	79
IT diplomates				34	36	36
Other engineering graduates				36	56	56
Non-engineering graduates				32	43	43
				158	214	214
Source: Nasscom 2002.

The software industry is labour-intensive. The essential work in it is done by human workers; without them it cannot grow. They have to be trained to write software; and most who have been so trained have been engineers. Hence behind any growth projection is implicit a daunting projection of manpower requirements.

As shown in the accompanying table, Nasscom estimates employment in the industry at the end of last March to have been 416,000; within three years it expects the demand to grow to 929,000 at least and 1.4 million at most. Thus it estimates the industry to require half a million workers at least, and a million if things go well – compared to less than half a million total employment this year.

In Nasscom’s view, the present annual supply of new programmers is 158,000; with planned expansion of the educational facilities it might go up to 214,000 by 2004-05. With the existing and planned capacity the stock of programmers may go up to 875,000 in 2004-05 – well short even of the minimum requirements. That figure itself is optimistic; it assumes, for instance, that the net emigration of 64,000 programmers in 2001-02 would be reversed into a net immigration of 8,000 by 2004-05.

What drives Nasscom’s stratospheric demand estimates is the demand from software exports, which it expects to go up from 180,000 workers last year to 641,000 at least and 1.06 million at most in three years. These export demand estimates are driven by four factors. First, the proportion of big American corporations hiring foreign software suppliers is projected to go up from 44 to 55-72 per cent. Second, their average IT budget would rise 5 per cent a year, from $196 million to $ 227 million. Third, the share of software services in their IT budgets is projected to go up from 65 to 75 per cent. Finally, the proportion of that share going to offshore suppliers is expected to rise from 12 to 25 per cent.

How realistic are the assumptions? At 44 per cent, the proportion of large US firms buying software services abroad is already high; but there is no reason to doubt that it can go higher. The question is whether the new clients are likely to be spending more than the old ones or less. My own guess is that those who went out of the US first were the big spenders, for which their big budgets made outsourcing a part of their needs less risky. Those that follow will be spending less; the average expenditure per company is more likely to go down than up. Suppose that it did not rise from the current $196 million; that alone would bring down the minimum demand below the projected supply of 875,000.

Will the proportion of IT expenditure on software go up? The assumption is that the big companies have the equipment they need and will spend more on getting more out of it. But as work requirements grow, there is also often an argument for junking old equipment and going for new. I suspect it will all depend on software. A major function of software, especially that sold by Indian firms, has been to get more out of old equipment and to make it work together with new equipment. If they get better at this, they will encourage retention of old equipment, and vice versa. It is best to be agnostic on this point and to assume that the proportion spent on software will remain the same.

It is the assumed rise in the proportion of software expenditure spent abroad that is the most uncertain. For it is a fact that American firms are outsourcing more, and more central IT-related functions. As they do so, they will be increasingly concerned about security, continuity and accessibility. In other words, local suppliers have an inherent advantage in outsourced services, and remote suppliers like the ones in India have a disadvantage. This does not necessarily mean that Indian firms will lose out on outsourced services. But if they want a share of this market, they will have to supply it from locations in industrial countries – especially the US. And then the number of Indian programmers they can use will depend on those countries’ visa policies. Germany is supposed to have introduced a green card in 2000 specifically to attract Indian programmers. When I went to get a German visa in May, I found that every young person was being given a standard letter saying that he or she was being denied a visa and that German law protected the embassy from having to give a reason. Even in the US, when Clinton increased the number of H1-B visas in 1999, Congress added a rider that any employer asking for an H1-B visa had to certify that he had tried to get the workers within the US and failed to get appropriate ones. Thus, even if more work comes to Indian firms, it will not necessarily be done in India, and may not even be done by Indian workers.

Finally, Nasscom assumes that the productivity of workers will remain the same. In actual fact, however, there is a progressive rise in their productivity. Firms copy already written programmes into the new work they do, and more and more shrink-wrapped tools are coming into the market that are designed to increase worker productivity. The software industry is still in its infancy; at this stage, enormous increases in productivity are possible, and will take place.

Thus even if Nasscom’s forecast of software expenditure were to materialize, it would generate far less employment. The demand for additional workers will be well within the supply of newly trained programmers, and there is no reason to expect continuing shortages or to argue for ever larger training facilities.

In fact, there is a chance that the demand will fall short of supply; in which case programmers’ wages, which have already fallen in the last two years, will fall further. They could well fall until they are no higher than those of other graduates. I also see no rationale for the software firms’ preference for engineers. The CEOs I have talked to agree with me that many engineering colleges are second-rate. But still they prefer taking engineers because, they say, engineering colleges select their students more rigorously and because engineering teaches logic which is essential to programming. But some of them have also told me that their best programmers had been non-engineers. I think software firms recruit engineers because they are headed by engineers. If they are in a bind and look around, they can find equally good potential programmers amongst non-engineers. In which case the supply of programmers will be even larger.

So, contrary to Nasscom’s projections of ever worse shortages, programmers may well become abundant and cheap; and therein lies the best hope of Indian industry. Low wages have been its greatest strength in the past; they may come to so in the future as well.

VII     The Key Issues

There is recurring debate in the country on the overseas versus the domestic market; many people feel that the industry should be selling more software at home. This was the inspiration behind the Task Force on Information Technology: it deliberately raised projections of domestic sales above and lowered those of exports below past trends. Those projections fell flat on their face: export growth since 1999 has exceeded the growth of domestic sales by an even greater margin, and the share of exports in revenue has continued to grow. But this yearning for domestic growth, or bias against exports, persists; it is an expression of the strong xenophobic and autarchic forces that have always been present in this country.

Related to the yearning for greater domestic focus, but different in its inspiration, is the view that the industry should serve the masses. It should help farmers sell their buffaloes and find grooms for their daughters; it must raise the awareness of the poor. These views come from the socialist strand, another persistent and powerful force in the country.

Without going into the merits of these opinions, I would only say that many people entered the IT industry with precisely these ideals of serving the country and serving the poor. They tried their best, but failed. The domestic market has consistently been less profitable and more difficult to develop; some firms persisted in developing it, but the export tornado swept them along. Any firm that wanted to survive and expand had to export, for that is where the market was growing, where profits could be made. I do not really see this changing in the near future: India’s economy is small and poorly developed, and has little capacity to use IT. It can be made to use more: the way to do it is by subsidizing domestic sales – or sales to the poor if you wish. The income tax concession on exports is slowly being withdrawn; it could be replaced by a similar concession on domestic sales. Alternatively, direct subsidies could be given to domestic sales of software. There is a good argument for it, not based on autarky and xenophobia: the industry’s success abroad depends on its domain expertise, and subsidy to domestic sales would expand this expertise with advantage to the country.

Another national obsession is with the training of software manpower. Manpower shortage has plagued the industry almost since its outset; consequently it has lobbied the government for the training of ever more programmers, generally with an engineering background. This lobbying is understandable; but as I argued in the last article, the manpower shortage is at an end for now, and it would be wrong to assume that it would recur. We have developed an untidy but effective system for training thousands of programmers quickly; it has worked in the period of the industry’s most rapid growth, and it will work in the future also. There is a case for reforming universities – giving them autonomy, increasing their material and intellectual resources and making them more responsive to the market. But this case is general, and not related to IT alone.

The real issue is, how much of the world IT industry will locate itself in this country? As I showed, a tiny percentage of the global IT market is served from India; and the markets served are confined to Anglophone countries and custom application software. There are vast markets in terms of applications and regions which India has hardly touched. It is customary to see in this a great opportunity for Indian firms to raise their market share; but within that figure is also the failure of the Indian industry to win a higher share, and the possibility that its market share might fall – in the extreme case, down to zero.

What I mean is that there is no inevitability about the industry’s location in India. The large Indian firms are already partially foreign – in their markets, in their sources of capital, in their partnerships, often in their workers. The bulk of their labour force as well as their legal location are in India; but both could move out. The IT industry is not tied to the soil, like tea or mica; it could fold up its tent and leave the country tomorrow. Even without doing so, Indian firms could employ more Chinese and fewer Indians; or they could warehouse Indian programmers abroad, as some Israeli firms are already doing. The turnover of IT firms run by Indians in the US was only slightly lower than the exports of firms located in India in 1998. The number of Indian programmers in the US is at least half the number in India, probably higher. These ratios could change either way.

Of the factors which will determine them, two are important. One is the internal and external security situation. India’s confrontation with Pakistan as well as the Gujarat riots did enormous damage to the Indian industry’s efforts to attract business from the west – especially efforts to get a share of the expanding BPO market. To get this type of sensitive business, India simply has to join the secure part of the world; and it cannot do so as long as it keeps publicly scrapping with Pakistan. This is not a judgment; I am simply drawing the conclusion from what many in the industry have told me. If we want sophisticated activities to locate and create employment in India, we simply cannot afford to continue our present internal and external security policies. China, in many ways a more autocratic and oppressive state, understands this; it has put all its external conflicts on the back burner and kept a tight leash on internal security. There is international competition in security management; it strongly affects the growth of sensitive industries like IT, and hence national growth.

The other is the threat of Dutch disease. If the software industry continues to grow at its historical rate, its exports will finance all India’s external payments for goods and services by 2008 and will exceed its GDP by 2013. That cannot happen: a part cannot exceed the whole. Much before that, either the exchange rate will appreciate, or prices in India will rise faster than international prices, or imports will flood in. Whichever happens, Indian industries will become less competitive and die; their decline has already begun. If domestic costs rise fast enough, the IT industry itself will become uncompetitive and leave India. Keeping it healthy and growing in India requires skilful macroeconomic management – far better than our government has shown itself capable of. There is not even a sign that the government understands the issue. The greatest threat to the IT industry is the quality of government.