American companies begin to give up R & D: who should pay for corporate research?

For a long time, American enterprises have been at the forefront of science and technology: companies such as general electric company, DuPont company and at & amp; T Bell Laboratories all promoted the great progress of the 20th century, and led the innovation and development of integrated circuits, plastics and synthetic fibers, which in turn became a powerful driving force for economic development.

now, things have changed. Are American enterprises still as innovative as they used to be? In a recently published paper, professors at Duke’s Fuku School of business argue that the shift from formal business research to a more decentralized innovation ecosystem driven by start-ups and universities has led to a decline in innovation capacity and economic growth in the United States.

mark discussed with Ashish Arora and Sharon belenzon, professors at Fuku business school, about the demise of enterprise research labs, how innovation works in new ecosystems, and why solving complex problems has become increasingly difficult.

Ashish Arora: in the early 20th century, American industrial research investment was small, unbalanced and concentrated in large companies. To some extent, this reflects the weakness of American university system. At the time, the American university system approach was very practical – serving farmers, small manufacturers, and miners in the States.

after World War II, we have entered the golden age of American enterprise research. At & T, DuPont, general electric, Kodak and many others have made large investments.

since the mid-1980s, the pendulum began to fall back, and many companies began to withdraw from research, with only a few exceptions. One of the most famous is Microsoft, which has built a very important research business.

Nowadays, there is very little enterprise research, except for the fields of artificial intelligence and machine learning, and to some extent, the field of biopharmaceuticals. There are some publicly listed biotechnology companies that are still conducting investment research, but in general, the U.S. corporate sector is in the long process of withdrawing from research.

However, in the broader ecosystem, American universities have done a lot of research and cultivated a large number of trained scientists, so that we can see a dynamic ecosystem with many start-ups. On the one hand, it clarifies the division of labor between universities and start-ups; on the other hand, existing companies can also use these inventions. But there are some gaps in this new ecosystem. Although this method works well in some places, it may be ineffective in some areas.

Ashish Arora: I think the trouble we have is more complex innovations, such as looking for new semiconductor materials. There may be startups working on this, but if big companies like IBM, Intel and Google don’t make significant investments, it’s hard to make a big breakthrough. Part of the reason is the scale, and part of it is the scope of the problem.

for example, the development of technological solutions to our complex climate problems requires hardware, chemical and electronic equipment, and more importantly, changes in regulations during implementation. Big companies need to constantly respond to regulation, but for a venture backed start-up, it’s very difficult to persuade national regulators to change the way they regulate equipment.

Sharon belenzon: what I disagree with Ashish is whether it makes sense for companies to link up with science. I think the answer is yes. Ashish prefers an ecosystem in which companies and departments are more specialized and more aligned with different activities in the innovation process.

but my interpretation of history is that only companies that have realized very early that to succeed, they must have a deep understanding of natural phenomena and thus be one step ahead of their rivals can be winners in the 20th century. What worries me is that if U.S. companies rely solely on other parts of the innovation system to develop the scientific basis of their business, they may lose their scientific edge and therefore their dominant competitive position.

American companies are gradually withdrawing from research, and I don’t think scientific research conducted by universities and small companies can make up for this. Leading economists in the industry believe that the unique feature of the United States lies in the close connection between technology and science produced by the diversified capitalist system. The diversity of institutions involved in R & D is key. I fear that we are losing that diversity in exchange for the greater efficiency that specialization brings.

Ashish Arora: there are two main factors. One is that there are many other sources of knowledge, such as university systems and start-ups; the other is that research is an unnatural activity within the company. The main purpose of the establishment of enterprises is to produce and provide goods and services to the market, rather than to allow activities without clear deliverable results to be carried out for a period of 4 to 6 years rather than 6 to 18 months.

in the golden age of enterprise research, DuPont developed nylon products and other early successful cases, which gained a lot of goodwill. As a result of these early successes, companies agreed to continue to inject capital. But at some point, the goodwill will disappear.

Microsoft is a good example. In a way, Microsoft research is a big company that invests in basic research because Bill Gates has a personal passion for it. After gates, Steve Ballmer was also the product of this original system, and then he carried it forward. But when Satya NADELLA was chief executive of Microsoft, Microsoft Research ended up as a large university system, because he wanted Microsoft Research to focus more on making new technologies available to consumers.

these things have a natural life cycle. Given the pressure faced by enterprises, it is difficult for a typical company to maintain research activities. At & T can invest in research for a period of time, because they are actually regulated monopolies whose research is part of social transactions, but after the monopoly stops, at & T will have to give up their laboratories.

so, do these large enterprises have the feeling that they don’t need their own laboratories because they can outsource their R & D work to the ecosystem and universities of start-ups?

Ashish Arora: we shouldn’t call it outsourcing, because outsourcing means I’ll give you a contract, and you do it for me, but that feeling does exist.

we should not call it “R & D”. Research and development are two completely different activities. Development is a standard activity with manageable endpoints, and research is the problem. Research usually accounts for 5% to 20% of R & D, so most R & D is development. Development is cool. In pharmaceutical companies, development means operating any clinical trial. Research refers to the study of the form of protein, or what the active part of the disease is, and then how this is transformed into a drug for Parkinson’s disease, which is the development problem.

in your paper, you talked about the difference between “technical uncertainty” and “business uncertainty,” and how it affects innovation. Can you explain it?

Ashish Arora: in the pharmaceutical industry, the main uncertainty is “will it work?” “will a molecule or vaccine work as expected?” this is mainly technical uncertainty. In digital applications, it’s not a problem to write an application, but the question is, “will people like it?” this is mainly business uncertainty.

our assumption is that when only one of the above two uncertainties occurs, the current system can still work well. But if there are two kinds of uncertainties at the same time, the situation will become difficult, and the current division of labor can not solve the problem of uncertainty in both dimensions.

how does this novel coronavirus outbreak play a role? Do you think the solution has the ability to solve both high technical and commercial uncertainties?

Sharon belenzon: in my opinion, we don’t have a sufficient scientific explanation for the behavior of different viruses, the way they spread, how they stop and how they interact in our bodies. Obviously, from an economic point of view, it is reasonable to invest any money in the development of such a scientific understanding, because this investment can hardly be greater than the huge economic and human losses we have suffered.

so the real question is, why haven’t we made this investment yet? Our innovation ecosystem has changed over the past three decades, making it more profit-making, and that’s why we need government. I am a strong supporter of the market, but also aware of the market’s shortcomings. What is good for individual enterprises may not be good for society – the market usually favors enterprises with strong practicability.

Ashish Arora: I agree with Sharon that science is difficult, but the biggest problem is business. Take the test or vaccine as an example, who is willing to pay for the test? Many people with risk diseases don’t want to pay for the test, but we want them to be tested, so now we need to solve this business problem by some means. I don’t think this is a problem that we have discussed and can be solved through enterprise research, but our system of guiding innovation through the market will lead to such a problem.

Sharon belenzon: This is a good point of view: the government is the number one consumer of new technologies, especially in the field of electronics.
think about the potential outcome of the former space race between China and the United States
, but I don’t think there is a possibility that a vaccine competition between China and the United States will bring about.

the government has provided a lot of support in the form of subsidies, but we have not yet clarified the difference between government funded research and downstream research and invention that actively participates in the market. Just giving you money to do whatever you want to do is different from starting a project that you know has a wide audience.

some historical cases show that when companies can rely on the government to buy anything they find, they can really take on huge scientific challenges, but the market itself is usually not enough to motivate people to solve very uncertain problems.