by Rakesh Mohan
India’s product and service markets have been transformed in the last twenty-five years. Has India’s innovation system been similarly transformed? If you are a macroeconomist, the answer is ‘no’. If you are a microeconomist, the answer is ‘maybe’, ‘somewhat’. In brief, the macro-innovation data shows no or modest change in the proportion of GDP spent on R & D, in who spends on R & D and in where it is done. There is some change in the sectors where R & D effort is concentrated. The import of technology shows rapid change. The individual firm story too is one of rapid change. In particular, a focus on learning has completely changed what products are made and how often new products are introduced. The basis of survival has dramatically changed the efficiency with which firms operate. Let us flesh out this picture of half-formation rather than transformation. I will focus on learning and R & D, leaving the discussion of changes in the higher-education system, trade policy, business and manufacturing to the specialized chapters in this book. I end with a discussion of what we must do to move from this half-formed National Innovation System (NIS) to one that is transformed. In particular, I argue that India’s unusual1 pattern of specialization in skill-intensive and capital-intensive manufacturing demands much more investment in innovation than currently happens.
1. A Few Building Blocks to Understand Technical Capability, R & D and Innovation
The importance of technical capability in economic growth is well recognized. From the very first growth-accounting exercises of the 1950s for the US economy,2 through analyses of Japan, South Korea and Taiwan in their catch-up stories, to China today, technical change is estimated to account for over half of all economic growth.3 Technical change shows up in the economy as ‘innovation’, as doing new things for commercial advantage. Innovation largely happens in firms, which must be at the heart of any such analysis. The innovative capacity of firms is affected by both what they do themselves and the institutions around them. The education system provides skilled labor, engineers and researchers. Where publicly funded research is done affects how it connects with industry. Public policy can provide incentives for investing in R & D, either directly or through patents. The trade regime can foster local production and/or an outward mindset. The culture of entrepreneurship affects investment in different kinds of capabilities. And broader cultural factors can influence how entrepreneurs define what is good. A framework of an NIS brings these factors together.4 It enables a more systematic comparison across countries of the drivers of technical capability. Although R & D is the most studied component of innovation, it is good to always keep in mind that innovation—defined as ‘something new for commercial advantage’—is a much broader concept and applies to all firms in all sectors. Innovation matters as much to a garment firm introducing a new design or a start-up launching a local-transport app (activities that rarely involve R & D as a formal activity done by dedicated technical staff, but are still highly innovative) as to a pharmaceutical firm developing a better cure for a disease involving years of research. R & D is, however, the most directly connected with the study of innovation, so we focus our analysis there.
1.1. R & D is Highly Concentrated
R & D is hugely concentrated worldwide. Most R & D is done in a handful of countries: of a total of around $1.5 trillion spent on global R & D in 2014–15, the top five countries accounted for 66 per cent, with industrial R & D at 71 per cent of the total.5 It is highly concentrated in a few industries: the top five industries—pharmaceuticals, automobiles, technology hardware, software and electronics account for 68 per cent of the total of industrial R & D. And within those industries, it is highly concentrated in a few companies: the top twenty companies account for 21 per cent of global industrial R & D, the top 300 companies for 67 per cent.6
1.2. India As an Outlier in R & D Spending
Table 1 shows that India was an early investor in R & D, with R & D as a percentage of GDP being higher than for countries that were considerably richer at that point in time. However, India’s investment in R & D has stagnated over the last thirty years, ranging between 0.6 and 0.9 per cent of GDP, while South Korea, Taiwan, Singapore, China and, to some extent, Brazil have substantially increased their investments. Mexico and Thailand (and Malaysia and Indonesia, as other leading NICs) reflect much more subdued investment in R & D throughout this period.
Table 2 shows that the bulk of R & D spending worldwide happens in firms (around 71 per cent of the total). The balance is publicly funded research, most of it done in universities (17 per cent), with a smaller share in autonomous R & D institutes (12 per cent).8 India is an outlier on three counts.
a) First, the share of industry in the total national R & D is the lowest of any major economy at 35 per cent.9 This share was 25 per cent in 1991, so the rise (of a rapidly growing GDP) is significant but not dramatic, and keeps India an outlier. The split of industrial R & D in India has changed significantly: in 1991, the 25 per cent industry share split into 15 per cent private-sector industry and 10 per cent public-sector industry; today, the 35 per cent industry share splits into 30 per cent private-sector industry and 5 per cent public-sector industry.10
b) Second, publicly funded R & D in India at 65 per cent is the highest among all major economies. China’s used to be high as well but, in this same period, has seen the publicly funded share fall to 16 per cent (from 50 per cent in 199111).
c) Third, where publicly funded R & D is done is again dramatically different in India. The bulk of public R & D (over 90 per cent) is done by the government in its own autonomous R & D institutes. A small share of publicly funded R & D (< 10 per cent of the public share) is done within the university system, giving the Indian higher-education sector the lowest share of national R & D (4 per cent) of any major economy.
1.3. Publicly Funded Research and Industrial Innovation
The popular mental construct of the relationship between scientific research and industrial innovation is simple: scientific research leads to discoveries that permit the development of new technology, and this new technology leads to into production and the market. This mental model, referred to in the literature as the linear model of innovation,12 is attractively simple but also simplistic. Over the last fifty years, the work of Kenneth Arrow, Paul David, Stephen Kline, Richard Nelson, Keith Pavitt, Nathan Rosenberg and Derek de Solla Price have greatly enriched our understanding of the true—and quite limited—role that scientific research plays in industrial innovation.
Price showed some fifty years ago that new scientific discoveries appear in industrial innovation with a typical lag of some twenty-five years. As such, an understanding of old scientific findings is adequate for most industrial innovation. This understanding of old scientific research is usually fully captured in course teaching, which leads one to the conclusion that science education matters much more to most industrial innovation than new scientific research.
Indeed, far from being the dominant source of industrial innovation, new scientific research matters globally to industrial innovation in just two exceptional cases. First, advances in certain fields, such as biotechnology and semiconductors, have close connections with scientific research. Second, there is a broader role for scientific research as one of ‘technology’s wellsprings’—to reinvigorate technical progress in a particular field.13 This ‘reinvigoration’ typically takes the form of a new technological paradigm for industry—online music taking over from compact discs, say, or the jet engine from the propeller. As Nelson puts it:
There is persuasive evidence that in many industries technological advance is what Winter and I have called cumulative, in the sense that today’s new technology not only provides enhanced operational capabilities but serves as a starting point for tomorrow’s efforts to further advance technology. Science may be involved as well, but in most industries science seems to be tapped as a body of general knowledge relevant to problem solving, with ‘new’ findings not playing a special role. Where new science is not partic
ularly important, a steady flow of newly minted scientists and engineers suffices to keep the laboratory adequately up to date with the world of public science.14
1.4. But What about India?
So research is critical to technical advancement in science-based industries and to the innovation of new technological paradigms. The results of research can be appropriated by other firms and indeed by other countries. Only when particular industries, such as semiconductors in Korea and Taiwan or cars in Korea, approach the technological frontier is there a case for scientific research itself, and hence for publicly subsidizing it. Scientific research should be seen, then, as the follower, not the leader, of industrial activity. Keith Pavitt made just this point:
… national technological activities are significant determinants of national economic performance as measured by productivity and economic growth. But what about the causal links between developments in national science and in national technology? Do they run from a national science base that creates the ideas and discoveries that the national technology system can exploit? Or do they run from the national technology system that creates both demands on—and resources for—the national science system? Our reading of the (imperfect) evidence … is that the causal links run from the national technology system to the national science system.15
1.5. Where Should Publicly Funded Research Be Done?
University Research As an End in Itself
Although India was an early investor in scientific research, this investment went overwhelmingly into autonomous scientific research institutions. The end result was that research would bypass the university system, a point long understood everywhere except India.16 For example, the Council of Scientific and Industrial Research encompasses thirty-seven laboratories employing 4000 scientists: assessments of CSIR’s contribution to Indian industry (its reason for existence) have shown little connection with industry. Any attempt to reform the Indian scientific research system that does not address this core issue of combining public research with teaching and not doing it in autonomous research institutions will be fruitless. This lesson is just not being learnt: the Eleventh Plan (2007–12) set up fourteen new autonomous institutes, and the Twelfth Plan (2012–17) proposed seven more. We could at least grandfather the problem and allocate incremental public-research funding to the higher-education sector.
Combining research and teaching will benefit both. The huge growth in higher technical education in India has all been at the undergraduate level; graduate technical education has stagnated. As the better institutes, and in particular the IITs, attempt to grow their graduate and PhD programmes, a shortage of qualified faculty is becoming increasingly acute. World-class graduate education requires that teachers do research, and unless there is dramatic growth in research, we cannot hope to have world-class graduate education. But the benefits from combining research and teaching would not flow just one way to teaching. Research would benefit too. Thanks to India’s early investment in scientific research, by the 1980s, it had achieved the levels of a medium-sized developed country in the primary measure of science output—publications in scientific journals. But this lead in publications did not show up in patents, often used as a measure of the output of technology research, where Korea and Taiwan have been the big new entrants.
Learning from Korea and Taiwan, the flow runs sequentially from industrial development to industrial in-house R & D to public scientific research.17 An industrial sector competing with the best firms in the world in increasingly sophisticated industrial sectors is a requirement for sustaining investment in in-house R & D, and strong in-house R & D is a requirement for sustaining investment in public scientific research of value to industry. It is only since 1991 that Indian industry has increasingly had to compete with the world’s leading firms. This has, in turn, driven investment in in-house R & D by specific Indian firms and industries such as pharmaceuticals. The more advanced technological sectors in Indian industry are now capable of utilizing, and therefore sustaining, investment in public scientific research. By combining this research with teaching, the Indian economy will get the primary benefit of doing research: the availability of trained researchers.
2. R & D in Industry
Table 3A shows data by sector for the top 2500 R & D spending firms worldwide (who account for over three-quarters of global industrial R & D spending). Note that of twenty-six Indian firms (against 301 Chinese firms and eighty South Korean firms), nineteen are in just three sectors—pharmaceuticals, automobiles and software, and India has no firms in five of the ten top R & D intensive sectors worldwide. Part 1 of the explanation of why industrial R & D in India lags is this absence of several sectors that are R & D intensive.
Table 3B: R & D Intensity (R & D as a Percentage of Sales Turnover) by Sector (2014–15)
Sector
Company
Reported R & D Intensity
Top 2500 Global Average R & D Intensity
Pharmaceuticals and Biotechnology
Dr Reddy’s Laboratories Ltd
11.8
15
Cadila Healthcare Ltd
9.5
Lupin Ltd
8.9
Cipla Ltd
8.2
Sun Pharmaceuticals Industries Ltd
7
Automobiles and Parts
Tata Motors Ltd
6.1
4
Mahindra and Mahindra Ltd
3.7
Bajaj Auto Ltd
1.7
Ashok Leyland Ltd
1.4
Maruti Suzuki India Ltd
1.3
Technology Hardware and Equipment
Zen Technologies Ltd
16.6
8
Genus Power Infrastructures Ltd
10.6
Astra Microwave Products Ltd
3.7
ITI Ltd
2.1
Bharat Dynamics Ltd
0.8
Software and Computer Services
Oracle Financial Services Software
7.1
10
Infosys Ltd
1.3
HCL Technologies Ltd
1.1
Tata Consultancy Services Ltd
1
Wipro
0.5
Electronic and Electrical Equipment
Electronics Corporation of India Ltd
3.6
5
Crompton Greaves Ltd
0.9
Bharat Electronics Ltd
0.5
Philips India Ltd
0.3
Industrial Engineering
Bharat Heavy Electricals Ltd
3.3
3
BEML Ltd
2.7
Escorts Ltd
2.2
TRF Ltd
1.8
Cummins India Ltd
0.6
Chemicals
Syngenta India Ltd
3
3
UPL Ltd
1.4
Hindustan Unilever Ltd
0.2
General Industrials
Hindustan Aeronautics Ltd
6.7
3
Titan Company Ltd
0.2
Oil and Gas
Oil India Ltd
0.7
0.5
Oil and Natural Gas Corp. Ltd
0.5
Reliance Industries Ltd
0.3
Indian Oil Corp. Ltd
0.1
Construction and Materials
Larsen and Toubro Ltd
0.4
1
Indian Hume Pipe Co. Ltd
0.3
VA Tech Wabag Ltd
0.1
Rail Vikas Nigam Ltd
0.03
Source: India Annual Reports (various years), Department of Scientific and Industrial Research (DSIR); Annual Reports (20
14–15) of Indian companies, EU Industrial R & D Investment Scoreboard (2015); Centre for Technology, Innovation and Economic Research.
Tables 3B and 3C provide Part 2 of the explanation: leading Indian firms invest somewhat less in R & D as a percentage of sales than their global counterparts. But a much more dominant explanation than the proportion of sales spent on R & D is the absence of really large R & D spending firms. No Indian firms, for example, figure in the top twenty-five R & D spenders worldwide.
3. What Would Transformation Look Like? A Comparison with South Korea and China
South Korea saw two transformations in the twenty years from 1970 to 1990. First, the proportion of R & D done by firms and the state essentially reversed. Second, the share of R & D in GDP rose strongly. Absolute investment by firms in R & D rose dramatically: a rising share, of a rising share of a rapidly growing base means a double multiple. The industrial share of total R & D increased from 13 per cent of national R & D spending to 81 per cent, at a time when R & D increased from less than 0.4 per cent of GDP to 1.9 per cent, during which South Korea was growing at 8 per cent a year. The same has been true in China over the last twenty years: the industrial share of total R & D almost doubled during the trebling of the share of GDP spent on R & D, while China was growing at over 10 per cent a year.
How has this transformation in industrial R & D happened? There is a double source. First, South Korea (as well as China) has seen substantial structural change in lead industrial sectors. Textiles and apparel, and food processing (low R & D-intensity sectors worldwide) have seen their share in industrial output fall. Automobiles, semiconductors, electronics and IT hardware (high R & D-intensity sectors worldwide) have seen their share rise. In India, automobiles are the only R & D-intensive sector to substantially increase their share of industrial output, a much more modest structural change.
