THE PERIL AND THE PROMISE

While in Stanford, I completed a study of the Indian software industry {http://web.stanford.edu/group/siepr/cgi-bin/siepr/?q=system/files/shared/pubs/papers/pdf/credpr70.pdf) in 2001. This is a summary  that went into columns of Business Standard.  The software industry put Bangalore on the world IT map, and made south India richer than the north for the first time in centuries. It had nothing to do with the 1991 reforms; but if we regard the reforms as a success today (leading to high growth uninterrupted by payments crises), that success has a good deal to do with the rise of the IT industry. The BJP government was friendly and supportive of the software industry, but it was pretty clueless on macroeconomics, to which  I gave it space in this series.

The Peril and the Promise

I   The Terrifying Prospects of Good Fortune

India’s software exports in 1999-2000 were expected to reach $3.9 billion – 6 per cent of its current account receipts. That is not much. But India’s richest man in 1999 (in terms of publicly known assets) was Azim H Premji, the principal shareholder of Wipro, a software exporter; Wipro had the largest market capitalization in the Indian corporate world.  K R Narayana Murthy of Infosys was No 4 and B Ramalinga Raju of Satyam was No 12 in the list; their software companies were also amongst the most valuable in the Indian stock market. Further, software exports had grown 51 per cent a year since 1993-94. Projection of this growth into the not-too-far future leads to startling conclusions about the share of software in the Indian exports and economy.

The sudden riches and the glowing prospects have focused public attention in India on the software industry. The Government of India set up a Ministry of Information Technology and appointed a Task Force to work out an Information Technology Action Plan. Its report starts with the following words:

The government of India, recognizing that the impressive growth the country has achieved since the mid-Eighties in Information Technology is still a small proportion of the potential to achieve, has resolved to make India a Global IT Superpower and a front-runner in the age of Information Revolution…

I share the belief of the industry and the political leaders that the industry will continue to grow rapidly, and that its growth can contribute to a break with India’s lackluster record of overall growth and human development. But there are different growth paths, some more promising than others; it is worth examining the path chosen by the establishment and considering alternatives. Specifically, the government, in collaboration with the software industry, has chalked out a strategy that is sure to fail: it will damp down the growth of software exports and severely limit the broader economic benefits resulting from it. To exploit the software success, India needs to make radical changes in its macroeconomic policies – before it is too late.

In an industry that is growing so rapidly, considerable fluctuations in the growth rate are to be expected. Actually, however, the growth of software sales, both at home and abroad, has been fairly stable: in the past six years, the annual growth rate of exports has fluctuated between 47 and 61 per cent, of domestic sales between 31 and 52 per cent, and of total sales between 43 and 54 per cent. These figures do not suggest an industry being tossed about in a turbulent sea of competition; rather, it is an industry that faces an enormous market but whose expansion is constrained by supply factors – principally by the output of reliable software engineers and their retention by the industry. As the industry grows, this constraint will continue to dominate. I shall discuss it later. But for the moment, I wish to work out the implications of assuming that the growth rates of the recent past will persist, and compare them with the projections of NASSCOM.

NASSCOM projects exports to grow to $50 billion and home sales to $87 billion in 2008-2009; these projections are adopted in the official IT plan. They imply annual growth rates from 1998-89 of 32 per cent for exports and 55 per cent for home sales  – just the reverse of the past rates, for between 1993-94 and 1999-2000, exports rose 51 per cent a year, and domestic sales 30 per cent a year.

NASSCOM gives no reason for the export pessimism or the domestic optimism. But a likely reason becomes apparent if one projects software exports at the past rates. If software exports continue to rise at 50.9 per cent per annum, they will reach $158 billion in 2008-09. If payments for visible and invisible imports continue to rise at 7.6 per cent – the average rate between 1990-91 and 1999-2000 – they will reach $138 billion. In other words, India could very well finance all its imports of goods and services from exports of software alone by that year. If nothing else changed, it would not need any other exports – or foreign investment. And if other exports continued and foreign investment kept coming in, India would have to increase its import intensity. If it did not, it is very likely to catch the Dutch disease. It is already possible to discern its first signs: as Figure 1 shows, industrial growth as well as growth of commodity exports have slowed down, while reserves continue to grow.

The software industry has grown so fast that small variations in its projected growth rates make a large difference to its volume. The implications of the growth rates of the recent past continuing into the future are brought out more starkly by Figure 2: they imply that software exports alone will exceed all current account payments by 2008-9, and GDP in 2013-14. The idea that the software industry may dominate the economy and the balance of payments daunts people – especially those in the government and the establishment who mistake conservatism for realism.

There is nothing to worry about if the emerging trends disturb the establishment; as reality unfolds, the disturbance will pass. But if, as a result, the establishment takes steps to thwart the trends, especially such an exhilarating trend as software exports exhibit, it would no longer be acting in the national interest.

Let me spell it out more starkly. This is what Arun Jaitley is reported to have said at a recent BJP meeting:

If there is one road map of the Vajpayee government, it is to free the country from the shackles that prevented the flourishing of entrepreneurship. We want Indians to do as well within India as they do when they go abroad. The government can no longer be involved in areas it is not supposed to. India missed the industrial revolution, and this government is committed not to let the IT revolution pass us by.

There are many other signs of the government’s friendship towards industry: its ministers’ attendance of industry events, their openness to requests from industry, the Prime Minister’s Industrial Advisory Council – and the steady rise in industrial protection. In the circumstances, it is possible that this protection, which is the cement that holds together the alliance of the party and industry, will hasten the onset of the Dutch disease and abort the software revolution.

It is also possible that the entire thrust towards the domestic market that is embodied both in NASSCOM’s projections and the recommendations of the Task Force on Information Technology emanates from a lack of confidence that the export trend can be maintained, and seeks to change the direction of the industry’s development in favour of the domestic market. Is it trying to do so? To discover this, we need to look more closely at the Information Technology Plan.

II   Information Technology Policy Framework

Two important documents have driven the government’s policy towards support of the software industry: the Report of the IT Task Force appointed by the Ministry of Information Technology, and the Report of the Subject Group on Knowledge-based Industries prepared for the Prime Minister’s Council of Industrialists. The two are very different. The IT Plan was prepared in a great hurry, and much of it was implemented – at least in form if not in substance – with equal haste. It consists entirely of nostrums: no analysis, no argument, no consideration of alternatives. The Subject Group’s report covers a broader area – although the Group would have liked to think in terms of the application of knowledge to all industries, it was really thinking in terms of information technology and pharmaceuticals – and it was more knowledgeable about the former than the latter. Whereas the IT Plan was blatantly designed to favour the IT industry, the Subject Group tried to take a fairer, economy-wide view.

The IT Plan is divided into three parts on software, hardware and a national long-term policy. They simply list points of action to be taken – 108 on software, 84 on hardware and 131 on long-term policy. A classification of these points is given in the accompanying table. It shows that there is some truth to the supposition that the Plan is directed towards creating domestic demand for information technology: 15 recommendations of the software plan and 13 of the long-term plan are designed to expand the domestic market. Most involve computerization by the central government and in institutions under its control. There are other recommendations that, though not directly directed at market expansion, would in effect do so – for instance, recommendations to set up government data banks and to spend government money.

But market expansion is not the primary thrust of the Plan. Almost as many of its recommendations are directed at improving the supply and quality of engineers. If the Plan has a thrust, it is towards getting the government out of the industry’s hair; almost a third of the recommendations call for legal changes and simplification and relaxation of rules. And these recommendations for debureaucratization and leniency are concentrated in the area of foreign trade and payments, accounting for almost half of them. Admittedly, most of them come from the Plan for computer hardware; this industry, which has suffered considerably from import liberalization in the past nine years, is not asking for protection, but for debureaucratization of imports so that it can compete more effectively in the domestic and export markets. On the whole, the thrust of the report is not towards growth within a protected domestic market; if at all, it is towards a level playing field for competing with the IT industry elsewhere.

The Action Plan is best read together with the Report of the Subject Group on Knowledge-based Industries (2000). There is considerable overlap between the two. The main areas in which both seek action are the following (those on which action is already completed, for instance the regulation and taxation of employee stock options, are omitted):

Labour laws: The Subject Group on Knowledge-based Industries sought exemption from a host of rules relating to labour: rules limiting the number of hours, specifying hours of rest and minimum leave, requiring that wages be paid in cash, that walls and staircases be whitewashed once a year, etc etc. The IT Task Force had little to say on this; presumably the government representatives ensured that this embarrassing subject was kept out.

Labour supply: The industry is concerned about the shortage of software engineers, and many of the recommendations are directed towards raising the supply and enhancing the quality of education. The Subject Group was concerned about improving all education. The IT Task Force had numerous recommendations on training – almost arguing that every child should become computer-literate – at government expense of course – even before it could learn to read and write.

Customs and import control: Import duties on computer hardware and peripherals are substantial, but software exporters can avoid them by setting themselves up in zones specializing in exports or by promising to export services worth a multiple of their equipment imports. Movement of equipment out of export zones is policed; and licences to import against export undertakings involve red tape in their issue. These bureaucratic restrictions should be removed. The Action Plan proposed a new entity called the Soft-bonded Unit. Essentially, it called for an end to all policing, and to reliance instead on ex-post audits to ensure that export undertakings have been met. This was because it covered both the hardware and software industries. The software industry more or less avoids customs and import control, so the Subject Group had little to say on these.

Exchange control: Exporting companies should be freed from exchange control in respect of investment and divestment abroad by Indian companies, allowable expenses abroad and the use of credit cards. At present they are allowed to hold dollar accounts, but there are many vexatious restrictions on how money in these accounts may be spent; these restrictions should be relaxed. The Subject Group was more concerned about exchange control.

Finance: Apart from more finance from government banks and term lending institutions, the industry wants them not to seek collateral in the form of fixed assets, and instead to fund IT companies on the basis of their turnover.

It is typical for lobbies to ask for special favours and for the government to grant them. But such special favours to particular industries are undesirable for four reasons.

First, they greatly complicate the laws. Every law gets cluttered up with a string of exceptions.

Second, they give increased discretion to the bureaucracy that administers those laws: it can quibble about whether a particular favour is due to a particular firm in the particular circumstances, the quibbling can lead to delays, those who suffer from it bring extraneous pressures – “influence” – to bear on the bureaucracy, and where the political establishment and the bureaucracy are prone to corruption, exceptions expand the scope for it.

Third, they are generally discretionary and unfair to the excluded industries.

And finally, they reduce the impact of the relaxations by excluding other industries which might have proved equally responsive to them. Hence it is worth asking, in the case of all demands of industry lobbies, whether the concessions should not be extended to everyone.

Perfectly general changes of regime are desirable in all the above areas; I shall explore in the succeeding articles what they might be. Both the reports were prepared by industrialists, who have no clue about macroeconomics. And it is macroeconomics that is likely to turn the software boom into a crash unless it is handled correctly; I shall describe what that handling involves.

Finally, even if there were a general relaxation of the controls that will become dysfunctional, it would not resolve all problems. It is in the nature of economies that there are conflicts amongst macroeconomic objectives, certain strategic choices have to be made, and no particular strategy is inherently and permanently superior. I would also try to point out this vestige of indeterminacy in optimum policies.

Recommendations of the Action Plan of the IT Task Force
		Part I	Part II		Part III	Total	Part I		Part II			Part III		Total
							Per cent
Laws		4		6	3	14	3.7	7.1			2.3				4.3
Company law		1		4		5	0.9	4.8							1.5
Other laws		3		2	3	9	2.8	2.4			2.3				2.8
Simplification of rules		11		27	18	56	10.19	32.14			12.98				17.34
Central government		5		5	16	26	4.6	6.0			12.2				8.0
Customs		3		20	1	24	2.8	23.8			0.8				7.4
Other		3		2	1	6	2.8	2.4							1.9
Relaxation of rules		9		21	15	45	8.3	25.0			11.5				13.9
Exchange control		4		12	5	21	3.7	14.3			3.8				6.5
Banks and financial institutions		1			9	10	0.9				6.9				3.1
Commerce ministry		1		6		7	0.9	7.1							2.2
Other		3		3	1	7	2.8	3.6			0.8				2.1
Favours		24		19	7	50	22.22	22.62			4.58				15.48
Central government subsidies		1		2	2	5	0.9	2.4			1.5				1.5
Increase in financing		5		3	4	12	4.6	3.6			3.1				3.7
Reduction in customs duty		1		7		8	0.9	8.3							2.5
Reduction in income tax		7				7	6.5								2.2
Reduction in excise duty		1		4		5	0.9	4.8							1.5
Other		9		3	1	13	8.3	3.6							4.0
Market expansion		15		1	13	29	13.9	1.2			9.9				9.0
By central government		7		1	13	21	6.5	1.2			9.9				6.5
By telecommunications		6				6	5.6								1.9
By armed forces		2				2	1.9								0.6
Training		16			12	28	14.8				9.2				8.7
Expansion		12			9	21	11.1				6.9				6.5
Standards and quality improvement		4			3	7	3.7				2.3				2.2
Greater competition		6		1		7	5.6	1.2							2.2
Other government support		18		1	42	61	16.67	0			32.06				18.89
New institutions					13	13					9.9				4.0
Supply of information		4			9	13	3.7				6.9				4.0
Greater expenditure		6			4	10	5.6				3.1				3.1
Enforcement of standards					8	8					6.1				2.5
Financing of research		1			5	6	0.9				3.8				1.9
Computer security and censorship		5			1	6	4.6				0.8				1.9
Other		2		1	2	5	1.9				1.5				1.5
Bullshit		6		9	21	36	5.6	10.7			16.0				11.1
		108		84	131	323	100.0	100.0			100.0				100.0

III   Labour laws

The changes called for by the Subject Group on Knowledge-based Industries and the Task Force on IT Action Plan fall into three categories:

(i)             Relaxation in rules on working conditions:  The Shops and Commercial Establishments Act 1961 and the rules framed under it limit working hours to no more than 9 a day and 48 a week, require a break to be given at least once every 5 hours, limit the total length of a working day to no more than 12 hours in a day, and require that hours worked beyond 9 a day or 48 a week must be paid for at twice the normal wage. (How a worker not allowed to work overtime can earn overtime may be a mystery to a lay reader, but poses no difficulty to lawyers). The hours of work and wage rates have to be displayed, as also the applicable minimum wage and dearness allowance (cost-of-living adjustment). Every worker who has worked more than 240 days must be given one day’s leave for every 20 days he works. Before he goes on leave, he has to be paid wages in advance for the leave period. Advances may not exceed two months’ wages with the Labour Inspector’s permission. Every establishment must whitewash all inside walls, passages and staircases with lime and paint all internal structural iron and steelwork at least once a year, and maintain a record of the dates of such whitewash and painting. Wages can be paid only in coins or currency. Registers have to be maintained to record attendance, the hours worked, overtime, leave taken with wages, dates of whitewash and painting, fines or deductions for damage or loss imposed on workers, and advances. 

(ii)           Permission to use contract and casual labour: Employers have sought to evade the rigours of the laws by employing temporary labour. Temporary labour is legally defined as employment lasting not more than 240 days out of a year. Thus employers both employ a worker and dismiss him within 240 days, often temporarily; or they engage a contractor to provide labour on contract. To prevent this, the government passed a Contract Labour Abolition Act, which made contract labour illegal in a wide range of circumstances. It also amended the Employees’ State Insurance Act 1948 and the Employees’ Provident Fund Act 1952 to ensure that an employer of contract labour, and not the contractor, became liable for paying their health insurance and provident fund contributions. The Task Force recommended that the Contract Labour Abolition Act should not apply to the IT industry, and that temporary status should be defined as 720 days in three years instead of 240 days in a year. It also asked that IT firms should be allowed to dismiss 10 per cent of the employees in a year without permission. The Subject Group asked that knowledge-based establishments should be exempted from the liability for the health insurance and provident fund contributions of contract employees.

The motivation of the Subject Group for asking to be exempted from labour laws is clear: for the Group also asks that Labour Inspectors should not keep dropping by to do the inspections done by other Labour Inspectors already, should not waste the managers’ time, and should not be so incompetent as to be mystified by computerized records. The real grievance relates to the corruption amongst Labour Inspectors, who reapeatedly visit software manufacturers in search of bribes.

However, the problems pointed out are not peculiar to IT or knowledge-based industries, nor do they deserve exemption just for being those. No employer should have to face them. There is a case for rethinking the entire framework of labour legislation enacted by the first government of independent India in the 1940s in imitation of similar legislation passed by the Labour government in Britain. The relevant legislation is the following.

      i.         The Shops and Commercial Establishments Act is outdated and needs to be repealed. Even industrial countries that had similar laws have repealed or relaxed them. What the law should insist on is that there should be a written contract between the employer and the worker specifying the pay and working conditions; it should help in the enforcement of such contracts, as well as of contracts between trade unions and employers, by law courts. The regulation of working hours, days, days of leave etc is unnecessary and counterproductive. Businesses would have much better chances of surviving bad times if wages, hours and work intensity were flexible.

     ii.         The Industrial Disputes Act makes retrenchment impossible without the permission of the relevant State government, which has been routinely denied. This single measure has had enormous distortive effects. Establishments have kept below the legal minimum (of 50 workers with power and 100 workers without power) to keep out of its purview. Employers have dismissed workers within eight months so that they would not become eligible for permanence. And they have hired workers through contractors who can legally or illegally evade the ID Act.

The solution must be looked for in minimum levels of retrenchment compensation that would free establishments from having to seek permission. Thousands of workers have been involved in the schemes of voluntary retirement and severance carried out by firms in the past ten years. It would be helpful if their experience is studied and summarized, and lessons drawn for a legislated set of rules for minimum compensation.

   iii.         The Employees’ Provident Fund Act: State-regulated pension provisions covered 34 million salaried workers in 1998. Another 13 million salaried workers, 166 million self-employed persons and 97 million casual or contract workers were uncovered. Some of these could contribute to a Public Provident Fund Scheme operated through banks, which also returned 12 per cent. The employees’ provident funds allowed withdrawals under very generous conditions; so although they imposed very high compulsory gross contributions, the net accumulation was very small.

The Expert Group for Devising a Pension System found the investment pattern very inefficient. It recommended that all the various schemes should be replaced by individual retirement accounts with stringent provisions against premature withdrawal, to be invested with any of six pension funds, each offering three funds of varying level of risk, freely transferable across fund managers and funds, and operated through any post office or bank. Thus, the direction of pension reforms is now clear; instead of giving special concessions to the knowledge-based industries, the government should proceed directly to general pension reform.

   iv.         The Employees’ State Insurance Act provides medical treatment to workers of the covered establishments. It provides the umbrella for massive fraud by workers, doctors and bureaucrats; except in Bombay and Calcutta it hardly provides any worthwhile medical cover. Its working needs to be studied in some detail. In this area, as in pensions, the direction is clear: we must move away from state-administered, rigid structures to a scheme which is financed by beneficiaries through defined contributions, which is independent of a job or indeed of employment, and which relies upon the provision of services by competing insurers.

IV   Worries about labour supply

Cheap supply of skilled labour is the bedrock of the Indian software industry.It is also the operative bottleneck. The industry is extremely conscious of it; the Task Force’s recommendations on it border on panic. Amongst its suggestions are the following:

(i)             All students, teachers and schools who want them would get computers from donations, multilateral funds, large-scale imports at bargain prices and bank loans. So would colleges, universities, polytechnics and hospitals. Universities would be networked to provide distance education by 2000.

(ii)           Indian institutes of technology and the Indian Institute of Science must triple their intake of IT students. New institutes of information technology must be set up. Information technology must be made compulsory for all degree courses and introduced in some schools called smart schools. Schools in some districts that have achieved universal literacy must achieve universal computer literacy. At the other end, some educational centers in the most backward northern states must be made models of IT-based education. The army would set up Information Technology, System Engineering and IT Security Institutes. Soldiers being retired from the army would be trained in information technology and sent off into the villages.

(iii)          An Institute of Computer Professionals would be set up as an accreditation body for degree programmes.

(iv)          The Indian institutes of management must set up courses in project management and software marketing.

The Subject Group on Knowledge Industries is even more ambitious: it aims at “restoring the glory of the Indian educational system through a series of measures aimed at building a learning society”. Its agenda is very diffuse, but the following are its major elements:

(i)             The government should stop funding institutions of tertiary education, and should divert the money to schools. Instead, they should charge commercial fees, and their students should be given subsidized bank loans. They should draw funds from companies, charitable trusts and international financial institutions. Privately funded universities should be allowed. They should be given operational and financial autonomy.

(ii)           Teachers’ salaries should be raised. Indians teaching abroad should be lured into them. Half the appointments should be tenured; the rest of the teachers should come from the relevant industry. Teachers should be encouraged to take consulting assignments. Their research should be funded by industry and become result-oriented.

(iii)          A national agency to set educational standards should be set up. The bachelor’s degree should be revamped to make it the basis for employment or academic specialization.

(iv)          Schoolteachers’ salaries should be raised. The syllabus for Bachelor of Education should be modernized; teaching methods should move towards raising students’ interest and involvement and encouraging them to discover, experiment and invent. English should form an important part of the curriculum.

(v)           School syllabi should be diversified to encourage streaming; children should be encouraged to specialize in vocations.

The Subject Group’s otherwise excellent suggestions have only a marginal relevance to the IT industry. The IT Action Plan is better focused on the problems facing the software industry. But even then, the connection between them and the solutions proposed is most fragile.

First, because the industry faces rapid labour turnover, the firms have the impression of a grave labour shortage. But the turnover only denotes a labour shortage; the magnitude of that labour shortage is quite uncertain. There are similar complaints of manpower shortage in the United States, which has been the subject of a series of official and non-official studies. The most meticulous one, by Freeman and Aspray, examines the various measures, including the rate of unemployment (which is less than half the economy-wide unemployment rate), permanent labour certificates issued by the Department of Labour, H1-B visas issued, and the rate of wage increase. It concludes that a quantitative estimate of the labour shortage was not possible.

It is natural for the industry to try to ensure that there will be an abundance of qualified people. If a surplus emerges, it will not hurt the industry; but it will be a national waste. The government was similarly hustled into creating a surfeit of engineering colleges in the 1950s. As a result, when industrial growth slowed down in the 1960s, a chronic surplus of engineers emerged. Even now, IIT graduates migrate in significant numbers into other occupations; many of them join the Indian Administrative Service, and many more go to the management institutes for further education.

Further, a limited familiarity with computers amongst millions of otherwise poorly educated students is unlikely to help the software industry. The labour supply problem of the industry is not one of quantity, but of quality. Young Indians are perfectly familiar with the prospects in computing, and have for some years been flocking to software training institutions. Those who can, get into the IITs and engineering colleges; but many more go to private classes. Their demand has been behind the success of the two major training companies, NIIT and APTECH. Most of them have taken secretarial jobs or joined the computer departments of companies and institutions. What prevents them from becoming programmers is the shortage of appropriate training facilities.

The only researchers who have probed the sources of software professionals in some detail are Arora and Arunachalam in a report they did for the Sloan Foundation. They estimate the annual output of engineering graduates at 104,000 and of Masters of Computer Applications at another 10,000. Their estimate of the output of informal training institutes is about 25,000. If these proportions went back some time in the past and all engineering graduates and MCAs went into software programming, their share in the current software work force should be at least 80 per cent. However, they also estimate that half of the software professionals did not have a degree in engineering or computer science. They found many chiefs of software firms who either did not use or were embarrassed about using informally trained people. That could also lead to their understating the proportion of such people they are using. This must mean one or more of three things. First, a high proportion – three quarters or more – of graduates do not enter the industry. Second, a high proportion of graduate entrants - leave the industry – mainly through emigration. Or, finally, the output of informal institutes has been underestimated.

My reading of these admittedly skimpy data is that a high proportion of the frontline workers in the industry has been trained informally. The picture is the same in the US industry; according to one estimate, 41 per cent of the software developers in the US industry were non-graduates, 45 per cent graduates, and 14 per cent post-graduates. Of the 60 per cent degree-holders, only 40 per cent (i e, a quarter of all workers) had a degree in software engineering. US output of software graduates is 25,000 – only about a half of the annual rise in software-related employment. Thus, graduation is not an essential qualification for entry into software writing. Twenty per cent of the graduates who enter the industry have specialized in some other subject than software.

V   Labour problems – a nuanced view

In thinking about the labour shortage in the IT industry, one should distinguish between four types of workers.

               i.         Conceptualizers: Their skill lies in translating a problem or a need into a blueprint for software to resolve it. Conceptualizers are generally the highest trained in the IT industry; they often hold doctorates. But their basic qualification is a flair for thinking algorithmically. Talent is important in this occupation; and since talent is not taught or transmitted, conceptualizers are highly prized and paid.

The Indian industry has relatively few conceptualizers, and hence specializes in the later stages of software generation. The problem is that there no course of education for conceptualizers. At some point some Indians will begin to learn the art; but as long as there is a shortage of conceptualizers, they will be easily lured abroad. There is little scope for organized action – least of all by the government.

              ii.         Developers: Developers take the conceptual plans made by conceptualizers and translate them into reality: they work out the programming inputs required to realize the concept, allocate the work amongst programmers, and manage the project until a suitable programme emerges. The demand for them does not come from the domestic industry alone. Programmers often travel abroad. There they pick up work on their account; on coming back they start their own firm. The rate of new business formation in this industry is extremely high. Further, a large proportion of the employees in this industry try to go abroad – chiefly to Silicon Valley – and the ones with greater experience and better contacts abroad are more likely to get a visa or a green card.

This problem is particularly serious for the Indian industry because its managerial salaries are limited by its low value added per employee. But even the industry in Silicon Valley faces the same problem, though to a lesser degree. Its solution to the problem is stock options – the returns to the employee depend on the eventual success of the project. But even that is not a complete solution; eventually, the options must mature, and at that point, employees will get the choice of leaving.

So a new solution has been worked out in Silicon Valley. Programmers form transient teams and hire themselves out to companies. The companies do not employ them; when the project is over, the team also disappears. This structure shows some signs of emerging in India; but it is inhibited by the labour laws. This is the reason for the Subject Group’s stress on the reform of labour laws, and it is justified

iii. Modifiers/Extenders: These are the foot-soldiers of the industry who write code, test programmes, modify them and adapt them to changing needs. When the industry talks of a labour shortage, it is primarily referring to a shortage in this category. Still, the industry has maintained a growth rate close to 50 per cent year after year, and has found workers to support it.

The problem, if any, is with the quality of the training. What the industry needs is a process of certification. But here, the solution that the Task Force proposes – that a professional institute should be given a statutory monopoly of conducting examinations – is misplaced. This is the solution adopted in accountancy and law; even in these relatively stagnant disciplines, it has worked poorly. An export-oriented industry like the software industry could hardly risk obsolescence arising from a monopoly in granting qualifications.

As it happens, the US industry is facing the same problem, but is trying to solve it in a different way. The major companies that produced packaged software – for instance, Microsoft, Cisco, Novell, IBM and Hewlett Packard – already run examinations to test proficiency in using their programmes. National Skill Standards Board, an official body, is working Education Development Center and with the major firms in the industry to work out voluntary certification standards for the industry. The major Indian firms should similarly try to work out standards that would be internationally acceptable and would at the same time fit into their own skill requirements.

(iv) Supporters/tenders: These are workers who do not actually create, test or modify software, but who run computer-related services within user firms, or help such firms as consultant trouble-shooters. They also include workers with computers in IT-enabled services. They do not need to be programmers; but they need to know the software they have to work with well enough to operate it and to solve minor problems that it may pose. Of the total IT-related employment of 2 million workers in the US in 1998, programmers were only 626,000; 1.2 million were “computer systems analysts and scientists”- that is, people who worked with computers but did not programme. A more recent manpower survey commissioned by Information Technology Association of America places total IT-related employment in the US at 10 million; the bulk of it is in technology support. The additional demand in 2000 was expected to be 1.6 million, of which 600,000 would be in technology support and another 300,000 in database development/administration. Only 220,000 would be in programming. The composition of demand in India is very similar; the bulk of the demand is in organizations that use computers, for workers who do not need sophisticated knowledge of IT. This demand is being met by the private training institutes, and will continue to be so met.

Finally, it should be pointed out that the software industry also faces a threat of brain drain. The difference in wages is the most important element in the migration. But competition is not equally severe in all segments of the labour market. Conceptualizers as well as designers are trained on the job, and their supply is less elastic. They are more often lured away to the US, and the ratio of their wages in India to those in the US is high. Modifiers, extenders, supporters and tenders are easier to train, their supply is more elastic, and their wage ratio is lower. Paradoxically, advances in information technology are opening up opportunities for services that use such workers – for IT-enabled services such as accounting, call service and medical transcription. Thus labour supply will constrain growth differentially for different sections of the industry; its overall growth need not come down, but its skill-intensity may. Far from “moving up the value chain”, the industry may find it makes economic sense to move down it.

Lower prices have been an important reason why it has secured a foothold in a faraway market despite the handicap that continuous servicing is important and proximity to the customer is an advantage. But the ability of the Indian software industry to quote lower prices, as well as to pay lower nominal wages, depends crucially upon the fact that the purchasing parity of the Rupee is greater than its exchange rate – that prices in India, converted at the current exchange rate, are considerably lower than prices in industrial countries. This is an advantage that India must strive to retain. Whether it can do so or not depends on monetary and fiscal management, which in a later article.

Skill requirements in the IT industry
	Conceptualizers	Developers	Modifiers/	Supporters/
			Extenders	Tenders
Typical occupations	Entrepreneurs	System designers	Maintenance	System consultants
	Product designers	Programmers	programmers	Customer support
	Research engineers	Software engineers	Programmers	specialists
	Systems analysts	Testers	Software engineers	Help desk specialists
	Computer science	Computer engineers	Computer engineers	Network installers
	researchers	Microprocessor	Database	Network
	Requirements	engineers	administrators	administrators
	analysts	Chip designers
	Systems architects
Qualifications
Doctoral degree	Frequent	Occasional	Occasional
Master's degree	Frequent	Frequent	Frequent
Bachelor's degree	Common	Frequent	Frequent	Occasional
Associate	Occasional		Occasional	Frequent
High school	Occasional			Occasional
Knowledge required
Information technology	Critical	Important	Useful
Business and industry	Useful	Useful	Important	Useful
Communication and
organization	Important	Important	Important	Important

VI   Customs and import control

Under the Uruguay Round agreement, India is committed to the abandonment of import licensing by April 1, 2003. But exporters will continue to be given access to duty-free imports of inputs and equipment. For this purpose, the government will continue to license duty-free imports; and trading partners will insist not only that such licensing should continue, but that it should tie exports and related imports as closely as possible.

Currently there are two types of such licensing arrangements. For inputs, permissible imports are tied, by means of officially defined input-output ratios, to individual import transactions. Exporters can either import inputs duty-free against individual export orders, which makes ordering of inputs extremely inefficient, or can claim duty drawback after exporting, which take months to obtain and often involve much corruption. Some flexibility was introduced into this transaction-by-transaction approach by the Duty Exemption Pass Book (DEPB), in which import duty is credited when exports are made and debited when imports are made. But the simplification is problematic since the implicit connection to transaction is still there, and the customs can suspend the working of a pass book on suspicion of misuse.

For capital goods, on the other hand, imports at zero or reduced duty are allowed against a commitment to export goods worth a multiple of the equipment imports over a number of years under what is called the Export Promotion Capital Goods (EPCG) scheme. Exporters can avoid both types of controls by locating themselves in export-oriented zones or their variants (export-oriented units and software technology parks). But flows of all kinds – goods, waste materials, manpower – across the border between the zones and the so-called domestic tariff area (DTA) are policed by the customs.

The policing is the source of most of the complaints addressed by the Subject Group and the IT Task Force. Of the 84 recommendations of the IT Task Force on Hardware, 20 relate to customs and import control. The most important proposal relates to the elimination of customs bonding.  The Subject Group proposes exemption of firms located in software technology parks from customs bonding. The IT Task Force on software proposes its replacement by an export obligation. The IT Task Force on Hardware proposes a new category of importer called a soft-bonded IT unit. It would have a minimum fixed investment (Rs 5 million for a small-scale unit, Rs 10 million otherwise), it would export 25 per cent of its sales, and if it cannot do so, the difference between its sales and its input imports (including imputed import duty) would be at least 18 per cent of sales. Such a unit would import all inputs duty-free, and would be under no obligation to export. But on whatever it sells in the DTA, it would pay import duties on the embodied inputs as well as excise duty. The policing of individual import transactions would be replaced by a legally enforceable undertaking and by audits based on accounts of input use, to be conducted by private auditors. Transfers of unfinished goods between soft-bonded units or between such a unit and its subcontractors would be duty-free; transfer of waste from such a unit to the DTA would attract import duty, but the unit could declare any value, down to zero, for the waste. In other words, a soft-bonded unit would pay such duty as it wishes to and an auditor approves.

Thus, the industry seeks an exemption from customs bonding for only itself, although bonding is just as vexatious to every other exporter; textile yarn units, for instance, accumulate mountains of cotton waste that the customs do not allow to be taken out without payment of exorbitant duties. And it proposes a remedy that would make it virtually impossible for the Customs to check whether the correct duty is being paid. Calculation of duty drawbacks requires data on physical quantities, whereas business accounts are purely financial. Additional accounts in terms of quantities could be kept; but they would be no simpler than the present pass book. Further, if the customs question an exporter’s entitlement to a duty exemption, the matter is now settled close to the transaction. If the customs question the entitlement after the annual accounts are prepared and submitted, the settlement is likely to be much delayed, and the data required to settle it are more likely to be buried in detail. Hence accounts-based calculation will be more complicated than transaction-based calculation, and is likely to lead to more protracted conflict between the customs and exporters.

However, an accounts-based calculation would be feasible for inputs if all inputs bore the same import duty, and all outputs bore the same excise duty. Then firms could, when calculating taxes on the basis of their annual accounts, take excise credit on the proportion of output that is exported, and import duty credit on the same proportion of inputs. If, as a special case, all import duties are abolished, then the credit due would be the sum of excise duty payable on exports as well as on the product of inputs and the ratio of exports to total sales. If a value added tax replaced excise, the calculation would be even further simplified: tax paid on imports would automatically reduce value added and thus be rebated, whilst exports can be directly exempted from VAT. Thus the ideal system would be one where there is no import duty at all and a value added tax is levied on imports as well as domestic sales. That is what India needs quickly to move to.

Manmohan Singh, while he was finance minister, made considerable progress towards reduction in import duties as well as greater uniformity. After he left in 1996, the number of rates of customs duties has continued to come down, but actual rates have gone up. P Chidambaram imposed a special additional duty of 5 per cent for revenue which was removed last year. Yashwant Sinha has reduced the number of duty rates, but in doing so has often raised duties. Under him, the government has also been liberal in yielding to business demands for anti-dumping duties. And as import licensing requirements have been removed in conformity with Uruguay Round commitments, they have been replaced by extremely high tariffs, especially on consumer goods and agricultural goods.

The process of reducing customs duties needs to be resumed and taken to its final conclusion, namely zero duties. The sequence may be the same as in the early 1990s: abolition of the highest rate and its reduction to a lower one in a series of steps. In the current political conditions – namely the close connections between the government and protection-seeking industry – such a process is impossible unless the exchange rate is used to give matching average protection. A devaluation would be odd when the balance of payments is so strong; but it is the political price that has to be paid to get industry’s acceptance for zero duty. Removal of import duty is the only way of ensuring that the government does not become a partisan player in the game industries play of agitating for competitive protection.

VII   Exchange control vexations

The official position is that the government has freed current account transactions from exchange control. However, it needs to ensure that capital account transactions are not disguised as current account transactions. So transactions over a certain amount need government approval. The limits are higher for businesses than for individuals; but they are nevertheless real. Controls on capital movements have also been relaxed; but prior approval is the rule. 

The IT Task Force as well as the Subject Group asked for relaxation of many exchange control measures for IT firms alone. Thus, for instance,

(i)             Reserve Bank of India allows exporters to maintain accounts in foreign exchange, and lets them make payments up to $25,000 without its approval. This limit should be removed. Lists of allowable expenses should be expanded.

(ii)           Reserve Bank expects export proceeds to be received in six months; if the delay exceeds six months, the exporter has to explain the delay and get permission for postponement from RBI. The Subject Group wants approval to be dispensed with, and replaced by simple quarterly reporting. 

(iii)          Reserve Bank is fairly liberal with banking transactions; it is more restrictive about the transactions it allows against credit cards. The IT Task Force would like to be allowed to against payment by credit card, and all payments that are allowed through banks to be allowed against credit cards.

(iv)          Reserve Bank allows firms to invest abroad p to $25 million or 50 per cent of export earnings over the previous years, whichever is lower. The Subject Group would like this limit to be removed. The Task force would like investment in kind to be allowed.

These and many other demands show that the software industry is chafing at exchange controls and would like to be exempted from them. So would every other industry. Whatever relaxations are made should be made for all industries.

If software export earnings continue to rise at the rates recently achieved, exchange control will no longer required for conservation of foreign exchange. Reserves have risen steadily, and that stage may well have been reached. The question that needs to be asked, therefore, is whether exchange control needs to be retained for some other reason than to conserve foreign exchange.

It was proposed in the previous section that there was a case for abolishing customs duties altogether. If that is done, the government may well want to retain control on the exchange rate as a means to influence the terms of trade. Exchange rate targeting requires a high level of reserves in relation to external payments; especially if the exchange rate tends to appreciate, it can be held down only if Reserve Bank buys off foreign exchange. If required actively to manage the exchange rate, Reserve Bank would also want some control on capital movements.

If Reserve Bank wishes to hold even larger reserves, it makes sense to liberalize capital inflows. The major categories that are relevant in this context are:

(i)             Portfolio investment in India by foreign institutional investors: These have been allowed to invest since 1991. The limit of investment by all portfolio investors in a company was limited to 30 per cent of its equity, raised later to 40 per cent with directors’ approval. The 40 per cent limit should be removed; the limit should be for the board of a company to decide.

(ii)           Portfolio investment in India by foreign residents: The 1991 reforms excluded personal investors, and they remain excluded. This exclusion is unnecessary. Portfolio investment by personal investors should be allowed provided they open bank accounts and use custodians in India.

(iii)          Foreign direct investment in India by foreign companies: In the 1991 reforms, foreign companies were allowed to invest up to 51 per cent of equity without permission in “priority” industries; investment beyond that level or in other industries required case-by-case approval. The list of priority industries was lifted almost without any change from the 1971 policy defining areas which foreign firms and firms belonging to big business houses would be allowed to enter. Most of these industries were stagnant and unattractive by the 1990s, and the liberalization of FDI brought in little investment. Much more foreign investment came in with case-by-case permission from 1997 onwards. It is now time to give up case-by-case approval and the variety of limits on equity holdings. If some industrialists oppose unrestricted FDI and the government is beholden enough to listen to them, that argues at best for a negative list of industries with a preannounced limited life, not exceeding five years.

(iv)          Borrowings abroad of Indian companies: The government and Reserve Bank between them keep a tight leash on these, for fear of a debt crisis. There is a limit on total borrowings in a year, and Reserve Bank also monitors the maturity pattern. They reached a peak of $3 billion in 1994, but have been insignificant since. Since their appetite for loans from abroad has proved limited, there is no need for restrictions on them any longer.

(v)           Foreign direct investment by Indian companies: Reserve Bank allows FDI by exporting companies without prior permission up to a certain amount, and with permission beyond that amount. There are few companies in India with the capital or the will to make investments abroad. It is now time to allow all Indian companies, whether they export or not, to make direct investments abroad.

(vi)          Equity issues abroad by Indian companies: The government was fairly liberal with equity issues abroad. About 70 companies issued equity abroad in the early 1990s. But the returns on them were so poor that there has been no market for further issues. The issues are already effectively limited by demand, and do not need to be restricted by policy.

(vii)        Foreign portfolio investment by Indian residents: Reserve Bank allows Indians returning from abroad to retain their portfolio investments abroad. But apart from that, portfolio investment is not allowed at all. The effectiveness of the ban is very doubtful. Even if Reserve Bank is not prepared to give full freedom of investment abroad to domestic portfolio investors, it should allow the sale of a wide range of foreign stocks, bonds and mutual funds in the Indian market. Indian capital markets are today amongst the most efficient in the world, and they are heavily underutilized. In the circumstances, trading of foreign instruments can help the expansion of this industry in which India should have a comparative advantage.

Thus apart from some temporary controls on foreign direct investment into India and portfolio investment abroad by Indian residents, there is little need for exchange control. If Reserve Bank is to control the exchange rate, and for that purpose needs to hold a minimum level of reserves, such reserve holdings must be achieved through fiscal and monetary policy. That is one of their roles. The other is to ensure that the software export boom is not stifled by exchange rate appreciation; and that the rise in export earnings is accompanied by maximum real growth and minimum inflation.

VIII   How to grow faster

The growth rate of software exports far exceeds the overall growth rate of real GDP. This cannot continue; at some point, the growth rate of software exports must slow down, or the growth rate of nominal GDP in dollar terms must rise. There are four ways in which it can rise:

(i)             The exchange rate can appreciate. That would make software exports less competitive and bring down their growth rate; it would also make other industries less competitive, reduce exports and increase imports. That would be uncomfortable for at least some domestic industries and would evoke calls for greater protection. We have argued that import duties and QRs should be avoided: in which case the exchange rate would be the only instrument of protection; the less it appreciates the better.

(ii)           The (positive) differential between domestic and international inflation may rise. Just like appreciation, it would make domestic industries less competitive, including software, and worsen the balance of payments. If inflation is demand-led, the rise in domestic demand could reinforce the worsening of the balance of payments.

(iii)          The growth rate of real GDP may rise. This would be the most desirable outcome. Thus, the policy question boils down to: what policies are needed to raise the elasticity of supply of aggregate output? This is the question dealt with in this section.

(iv)          Import-intensity may rise. This requires a reduction in the cost of importing and of transporting imports inland. This is also dealt with in the next section.

Macroeconomic policies to increase the aggregate elasticity of supply boil down to three things: increasing the domestic savings rate, increasing the efficiency of investment, and increasing investment in infrastructure industries – or more generally in industries which have been a preserve of the public sector in India and whose growth is handicapped by poor institutional and legal structures. In this article I shall show how to raise the savings rate.

 Domestic savings rate: India’s savings ratio has been in the region of 20-25 per cent for the past 20 years. It has been significantly lower than in the rapidly growing economies of East Asia (except Indonesia). As the accompanying Table shows, over 80 per cent of the savings continue to be “household” (I e, non-corporate private), and 40 per cent of the household savings are in physical form – that is, in the form of physical assets or property.

A major factor that depresses the Indian savings rate and distorts investment is the tax system. It is not the tax rates; after the reductions in the 1990s, they are today generally lower than in industrial countries, and not much higher than in comparable developing countries (except for import duties). It is the tax collectors, who have a reputation for harassment and corruption. The number of income taxpayers in India is minuscule compared to the number of people whose income is estimated to exceed the taxable limit. Finance ministers throughout the 1990s have tried to raise the number of taxpayers by decreeing that persons with certain characteristics denoting wealth – a car, a telephone, a trip abroad etc – must file tax returns; as a result, the number of taxpayers has risen considerably. But it is doubtful that they will stay in the tax net. The fact is, that Indians fear tax officials like the plague, and would do anything to stay out of their records. The fear of the tax official leads to a preference for consumption over savings, for consumption leaves no trail. It also leads to greater investment in property and housing; transactions in real estate typically involve half the payment in cash and half by cheque. It inhibits investment in the production of goods and services.

Thus, an honest revenue service is the best way to stimulate savings and direct them into productive investment. However, a clean revenue service in the middle of other forms of corruption is not realistic. Elimination of corruption in public life, therefore, becomes essential for non-inflationary growth.

There can be a number of ways to remove corruption. For maximum effect, however, there would have to be reforms affecting the political system as well as the bureaucracy, and both the senior and the junior bureaucracy.

For the political system, a changeover to a proportional representation with a high cut-off point for entry into all legislatures would be desirable. Proportional representation would strengthen the party vis-à-vis the individual politician and shift political funding from the latter to the former. The cut-off point should be set so high as to ensure that at most two or three parties would get into a legislature. That would increase the probability of a party coming to power and increase its average tenure in power. Parties would then acquire a stake in good government, and would be less tempted by the chance of using the government to make money. The incentive to get legitimate party funding can be reinforced by giving parties state subsidies in proportion to the finance they raise.

The gazetted civil services are at present recruited from young men and women in their twenties by means of competitive examinations. The age of entry should be raised to 30, and prior work experience should count in recruitment. The examination, which at present can be taken in an absurdly broad range of subjects, should be confined to those subjects that are required in administration – English, law, accountancy, economics and mathematics. The present multitude of services should be unified into one. Salaries at all ages should be comparable to those in the private sector. Appointment in the civil service should be on contracts of 10 years at a time, renewable twice at most; and recruitment at the age of 40 and 50 should be open to mature people from the private sector as well. Shorter contracts will permit the number of officers at each level to be adjusted to requirements. At present, officers recruited in their twenties retire at 58. Consequently there are too many old officers, and posts are created or appropriated to employ them. Once entry and promotion bars are introduced at 40 and 50, the civil service can be thinned down at higher ages to match the required command structure.

The junior civil service, consisting of clerks and assistants, is at present chosen on the basis of examinations that have become a farce. They are appointed for life, and their promotion opportunities are negligible. This is the major cause of low-level corruption; it also results in moonlighting and low productivity. Clerks should be recruited on short-term contracts, not exceeding 5 years; and they should not be retained beyond the age of 30. At that age they should either take the entry examination and join the gazetted civil service, or leave government and go into the private sector. The absence of secure career prospects will ensure that the government will get young junior employees who either want administrative training before taking employment elsewhere or who are bright enough to have a good chance of becoming gazetted officers. Junior civil service salaries also should be comparable to those in the private sector.

Saving and Investment Rates
(as percentage of GDP at current market prices)
	1993-94	1994-95	1995-96	1996-97	1997-98 @	1998-99 *
1	2	3	4	5	6	7
Household Saving	18.4	19.8	18.5	17.1	19.0	18.5
Financial	11.0	12.0	8.9	10.4	10.4	10.9
Physical	7.4	7.8	9.6	6.7	8.6	7.6
Private Corporate Saving	3.5	3.5	5.0	4.5	4.3	3.8
Public Sector Saving	0.6	1.7	2.0	1.7	1.4	negligible
Gross Domestic Saving	22.5	25.0	25.5	23.3	24.7	22.3
Gross Domestic Capital	23.1	26.1	27.2	24.6	26.2	23.4
Formation
@ Provisional Estimates * Quick Estimates

Source: Reserve Bank of India (2000).

IX   Getting more out of resources

Last week I discussed how our low rate of savings could be raised; now it is necessary to ask, how we can use those savings effectively.

Efficiency of investment can be increased by intensifying competition in the financial markets. The accompanying Table shows the composition of financial flows. The figures extend only up to 1995-96, but there was little change over the previous 10 years, and there is no reason to think that the picture has changed significantly over the past four years.

The first point to note is the high proportion of financial investment in currency and deposits. To keep down the cost of government deficit finance, Reserve Bank normally plans on growth of money supply of 16-17 per cent a year, which translates into real growth of about 7 per cent, inflation of another 7 per cent and an increase in real money balances of 2-3 per cent. The numbers vary slightly from year to year, but inflation is built into policy. It cannot be reduced unless growth of money supply is cut down significantly.

The other important feature is the predominance of non-intermediated fixed-interest flows – loans and advances, small savings, life insurance funds and provident funds – which account for almost a half of the total fund flows. India has a well developed capital market, which has undergone considerable technological improvement in the 1990s; but in fact it intermediated at most 10 per cent of the flows. Of these, small savings, life insurance funds as well as provident funds could be mediated by the market if the government liberalized restrictions on investment. So could a significant proportion of loans and advances if bill markets were developed to finance local businesses.

Finally, a very high proportion of the flows is preempted by the governments: government securities, securities of banks and financial institutions (most of which belong to the government), small savings, life insurance funds and provident funds – which together come to over 20 per cent of the fund flows in most years – go to the governments. These funds could go into productive investment if the governments ceased to borrow except for productive purposes. Thus elimination of the fiscal deficit becomes important for improving the efficiency of investment as well.

Successive governments have moved towards a uniform central value added tax by reducing the number of rates. But this is not enough; the present transaction-based tax needs to be replaced by an accounts-based tax. Once it is done, the collection of value added tax and income tax could be unified. In other words, taxes would be levied on two bases, both calculated from annual accounts – value added tax and tax on business profits. The tax on profits leads to considerable evasion as well as avoidance, which can be eliminated by replacing it with the value added tax. There would then be only two major taxes: a tax on value added by businesses, and an additional tax on personal incomes above a certain limit.

The same principle should be applied to state sales taxes: the multiple-rated taxes should be replaced by a single, uniform value added tax, which could then be collected together with the central value added tax by means of a surcharge. The states should, if necessary, be compensated by being allowed to impose their own surcharges on personal income tax, which may be allowed to vary from state to state. Different statewise taxes on goods and services militate against a national common market; but there is nothing undesirable about differing rates of taxes on persons. This is impossible under the present Constitution; but it can be provided for in the changes about to be suggested by the sitting Constitution Commission.

Investment in industry will go waste unless infrastructure is at hand to support it. An increase in investment in infrastructure industries requires their privatization and management for profit, together with introduction of competitive structures where possible and regulation where free entry and exit are not possible. In all the industries, the latter segments can be separated: transmission lines in electricity, pipelines in oil, wired network in telecommunications, the ports in shipping, the airports in air transport, the roads in surface transport and the track in railways. These segments should be made subject to the common carrier principle and price regulation; the rest of each industry should be opened up to competition. 

Of these, the transport infrastructure contributes significantly to the domestic cost of imports, and an increase in its efficiency will raise import intensity. Thus, it is well known that Indian ports constitute a bottleneck and yet utilize only a fraction of their capacity because of a ban on the use of non-union workers. Mechanization, together with round-the-clock use of the hauling equipment, would increase port capacity manifold. The capacity shortage is particularly serious in deep-water ports of which India has few; bulk cargo is hence transported to better run neighbouring ports such as those of Colombo, Singapore and Dubai and reshipped in small vessels to India. More intensive use of deep-water ports could eliminate the reshipment costs.

The railways massively subsidize second-class and suburban passengers at the cost of freight. Successive railway ministers have saddled them with enormous surplus manpower. The railways need to be privatized. Railway tracks should be hived off into a separate corporation which would rent out the use of fixed facilities – tracks, sidings, stations and yards. The railway services can be opened up to competition by dividing up the rolling stock amongst a number of companies.

The same principle of separation of transmission lines and supply can be applied to power. Although the distribution networks must remain regulated monopolies, they can be divided up into a large number of urban or regional units, each of which can compete for power. It is possible to demerge the state electricity boards into a large number of power suppliers. Both the generators and the buyers must be able to operate in a single national market, mediated by the national grid. A competitive, unregulated wholesale market for power would also give an easier entry to cheap hydroelectric power from neighbouring countries.

In oil and gas as well, the emerging pipeline network must work as a common carrier network. It would bring down oil transport costs, and make the entry of new refiners easier. It would also reduce fuel costs with the replacement of expensive domestic coal by imported oil and gas. Oil- and gas-firing equipment is cheaper than coal-firing equipment; freer availability of oil and gas would stimulate wider use of energy and mechanization.

Increase in capacity and cheapening of transport and transmission would not only raise import penetration, but it would also, to a much greater extent, stimulate domestic trade. Thus today, railway freights are so high that it is cheaper to transport marble to Bombay from Italy than from Rajasthan; cheaper domestic transport would improve the ability of northern producers to compete in the expanding markets along the coastline to the west and south.

Flow of funds by type of instrument (per cent)
	1986-87	1989-90	1992-93*	1995-96**
1.Currency and Deposits	30.9	28.5	26.7	23.1
2.Investments	24.9	23.6	26.2	20.4
a.Central & state govt securities	17.1	12.7	8.6	10.7
b.Other Government Securities	1.7	2.2	2.2	0.1
c.Corporate Securities	2.5	3.2	7.7	9.1
d.Bank Securities	0.9	0.5	0.4	0.6
e.Other Financial securities	3.6	5.7	5.9	2.2
of which
i) Mutual Funds	0	2.6	0.9	-0.4
f.Foreign Securities	-0.9	-0.8	1.4	-2.3
g.Others	0	0	0	0
3.Loans & Advances	23.7	26.3	29.4	37.8
4.Small Savings	3.6	4.9	2.1	2.1
5.Life Insurance Funds	2.4	2.7	3.4	3.2
6.Provident Funds	5.5	5.8	7.1	5.1
7.Compulsory Deposits	-0.3	-0.1	0	0
8.Trade Debt or Credit	1.2	-0.4	-0.2	0.2
9.Other Foreign claims	0	-0.1	-2.4	0.4
10.Other items not elsewhere classified	8.2	8.7	7.9	7.7
T O T A L	100	100	100	100
*Provisional Estimates
** Tentative Estimates