India has arguably the world’s fastest growing startup ecosystem. Yet, none of our companies made it to the MIT Technology Review’s Smartest Companies list (2012-2016). In contrast, China had 14 and US had 105 companies on that list. One of our companies made it to the list in 2017, but I do not see any great globally competitive technological innovation from the firm. In the same vein, 106 of the world’s unicorns are in the US, 98 in China and only 10 in India.
Why don’t our companies create global impact and significant economic value? We do not have the world’s largest internet companies, mobile companies, app companies and will probably not have the world’s largest artificial intelligence (AI) companies if we do not put our act together. Why couldn’t we take a leadership role in these disruptions? Research developments in universities and those funded by government led to these disruptions. An ecosystem of startups and companies in proximity to these universities, in geographical and other terms, made use of these disruptions to build large companies.
Google is a great example – it was born out of a Stanford PhD thesis. Equally good examples are the Silicon Valley buoyed by the California University ecosystem, led by Stanford University, and the Boston life sciences ecosystem buoyed by MIT and other great universities in the area. Similarly, the Chinese startup ecosystem is driven by their great research universities that took global leadership positions in the last 20 years. For instance, Tshingua University publishes more AI papers in top conferences than the whole of India, every year. Which great research universities are at the center of our startup ecosystems in Bangalore and Gurugram?
Our research universities and productive research manpower are few in number. Our total number of productive researchers is a mere 27,500, seven times lower than China and 17 times lower than USA. Another tactical mistake is that the bulk of our researchers (75 percent) work in government labs and not in universities. In contrast, in the rest of the world the ratio is skewed the other way around – the primary efficient way of research being faculty working with PhD students. Lack of research in universities, also means lack of good PhD programmes and choking the supply of future researchers. As a result, our high impact research output is just seven percent of US and 15 percent of China.
Research is a lacklustre career in India. The JEE and CAT are the talk of every media article. Who cares about GATE or NET, exams for pursuing a PhD in top Indian universities? Students perceive that the research career pays low, takes too long to complete, has poor professional outcomes and low overall social standing. While some of this is indeed true, our students are just not aware of some of the improvements over the last decade. Compare this with a PhD admit in MIT, Stanford or a UC Berkeley, that inspires awe among the best students of the country. Not only do we lose our top undergraduates to other countries but also to the industry. The bright ones who pursue research in India are only those in deep love with India or those who continue to be here for family reasons – not really attracted by market forces.
Thus the first step for India is to have many more research positions in top universities and to make the research career attractive. Then, we need to provide a liberalised and competitive research environment to the research community. Our universities work like pre-liberalisation corporate India– slow, bureaucratic and socialist in approach. Universities have little financial autonomy – freedom to spend money, faculty salaries, fee and promotion policy. Buying equipment and getting research funding is also marred by bureaucracy, delays and cronyism at times. For instance, to buy something that costs more than $5000, the process of tenders, committees, etc. takes three to six months as compared to overnight shipping in the US! 61 percent of our researchers at top universities rate getting equipment and funding in a timely and non-bureacratic manner as below average.
Liberalisation and autonomy has to be counterbalanced with competition and meritocracy. This is pretty much missing at all levels in our research ecosystem. It needs to start at the government, which should make its universities to compete for money rather than doling out money equally to them or according to whims of few influential people. Same stands true for research faculty. Low performing faculty do not have much financial disincentives. On the other hand, high performing faculty neither get financial incentives, nor the sway to hire more faculty in their are get more resources and decision power to build a large and successful program. Similarly, funding agencies generally prefer to fund more number of researchers with low budgeted projects than fund bold expensive proposals by high performing researchers.
By contrast, China is very effective in offering resources and incentives for performance. They carry out detailed benchmarking of individuals, departments and universities and accumulate resources to the high performing ones. China has been able to capture nine spots in the list of world’s top 200 universities in the last two decades. Universities in the US have the tenure system. In some cases, top universities show the door to 50 percent of their lower performing faculty members after the first seven or eight years. On the other hand, high performing faculty become celebrities within and outside the universities and gather resources around them. Consider, Rodney Brooks, Andrew Ng and Geoffrey Hinton – all celebrity figures in AI, with money chasing them! We also need to accumulate our resources in performing institutions and researchers, so that they can multiply their success and disrupt. Those not performing should have disincentives. Nothing can be worst for research than a socialist non-competitive environment.
The third issue is around collaboration. Researchers multiply their success by discussing their research, learning from each other and eventually collaborating. Unfortunately, the modern means of communication, the phone and the internet, are not good at this – they do not transmit ‘tacit’ knowledge and spur collaboration. Studies show that research collaboration decreases exponentially with the geographic distance separating researchers. We do not have the critical mass of researchers, the right structures and a culture of collaboration to facilitate such discussions. For a given field, say AI, number of our productive researchers, are in lower double digits. This also means that they have limited impact on industry and startups to create economic value.
On top of this, our interaction with the outside world is limited. Lack of travel money is cited as one of the top issues by researchers I spoke to. A faculty gets money for two international trips (plus an option to apply for one more) in three years. A PhD student gets money for a single international trip during her PhD. Both low and high performing researchers get the same amount of money. In contrast, the editor of a US technology magazine told me that he sees top AI researchers from China in the US almost every other month! I made six international trips in my two years at MIT as a graduate student. Thus, we do not have enough opportunity to learn and collaborate with international researchers. With no critical mass within and the lack of contact with the outside world, we become ‘frogs in a small well’!
I will now like to get back to where I started – how our research can fuel our industry and startup ecosystem to create globally competitive organisations. Today, there is lack of what I call ‘problem proximity’ – the proximity of our researchers to the real-world problem they seek to solve. Pre-liberalisation our industry and economy wasn’t mature, wealthy or large enough to interact with our researchers. Our research universities thus didn’t build the structures and culture to interact with industry. They didn’t access original questions from industry – rather they did derivative research inspired by Western papers. This made their research less impactful. At the same time they couldn’t and didn’t make a difference to the industry. This hasn’t changed even as our industry has grown manifolds in the last 25 years.
To realise the Indian startup dream, we need a high quality research ecosystem vigorously interacting with entrepreneurs, industry, venture capitalists and actual users of products. Industry needs to supply problems to researchers and researchers need to supply technology advancement and breakthrough solutions to industrial problems. Only a critical mass of researchers, who become entrepreneurs, consultants and manpower for the industry can bring about a change. A few would just fail and not move the needle. Over and above, we need a society and funders who respect ‘science entrepreneurship’, are ready to invest in them and go in for the long haul to create real value. Then only we can achieve ‘startup India’.