An Overview of Economic Security (5): From an Industrial Policy Perspective

Linear growth vs. exponential growth

DARPA, the Defense Advanced Research Projects Agency, is a key innovation center for the U.S. Defense Department. A presentation with similar content was delivered seven years ago, in 2016, to convince Toyota Motor Corporation (TMC) to establish the Toyota Research Institute (TRI) in Silicon Valley.

Standard Kaizen to develop ever-better cars and achieve continuous improvement was already happening in Toyota, as in many other Japanese companies. Looking at the example of the Toyota “Corona,” the progression of sophistication of this car model from 1957 to 2002 essentially shows linear improvement, and does not represent a discontinuity or disruption. On the other hand, in the area of cellphones, an incredible disruption occurred over the course of 30 years, especially after the launch of smartphones. The evolution was exponential. At the beginning there were small improvements, and then, over the past years, improvements accelerated tremendously. Because of Moore's Law, this applies not only to smartphones but to computer and IT technologies in general.

Linear progress achieved by the standard Kaizen process cannot compete with such exponential growth. This was a problem for Toyota and approaches other than the usual Kaizen needed to be found.

The background of DARPA

The creation of DARPA is linked to the history of the competition between the United States and the Soviet Union. In 1957, the Soviet Union launched Sputnik, the first man-made satellite. This disruptive advance of our competitor led President Eisenhower to found DARPA. The purpose was specifically to address such disruptive changes, which are also called “strategic surprise,” through research and development.

One of DARPA’s first activities was to work on the F-1 engine, one of the world's best rocket engines at the time. Only nine years later, the Saturn V rocket completed a successful flight and two years after that landing people on the moon was achieved. Other major accomplishments of DARPA include the Global Positioning System (GPS) and ARPANET which led directly to the development of the Internet.

Strategic surprises in the automotive industry

Seven years ago, automated driving efforts were undertaken by Google, Mobileye, and Uber and there were rumors about Apple working on car designs, too. These disruptive developments are a significant threat to established players in the respective industries and are “strategic surprises,” similar to Sputnik.

At the time, young people were becoming more obsessed with their cellphones than their cars. People interact with their cellphones for almost the entire day, while cars are only used at certain times. This trend is even more accelerated in today’s aging society. Mobility as a Service (MaaS) was another disruptive threat, which led to the reduction of brand loyalty.

Furthermore, the differences in product features between carmakers are much smaller than in the past and the industry is essentially mature. There are many competitors. In industries with a large network effect, bigger company size leads to being more competitive, resulting in a market with one or two gigantic players in a “winner-take-all” scenario. This is not the case for the automotive industry. Even Toyota, the top revenue company, had only around 12% of the global market at that time. By contrast, Google, which has a huge network effect because of the accumulation of data improving the quality of the results, accounts for over 85% of the search market. IT companies also have significant advantages in technology, infrastructure, their personnel, and finally, in the kind of glamor that is associated with working for them.

Updating the traditional Kaizen approach

Traditional Kaizen has been effective for improving products and procedures to achieve the best design. But what happens when market shifts occur, for example changes in consumer preferences? The old design will no longer be the best design. If the market environment changes slowly, traditional Kaizen can then slowly develop the best possible design over time by experimenting and finding better car designs.

However, if a paradigm shift occurs and the public desires a completely different kind of product that the industry is not aware of, and there is uncertainty about whether to switch to the new product or not and different opinions regarding the prudent course of action how should the company proceed? The new product may in fact be a better business than the existing business. Getting over this prediction uncertainty and having the courage to take a leap from the old product to a new product is the fundamental problem with the traditional slow Kaizen process.

Exploring each direction of possible new products may result in failure, which is very frightening for established companies. In many Japanese companies, particularly those with a long company tradition, there is a fear of failure that results in risk aversion. So how can the company proceed?

Lessons from DARPA

Every possible choice includes prediction uncertainty. What I proposed was that at the newly established Toyota Research Institute (TRI) our mission was to try all of the different directions and not to be afraid of failure, in order to find which new products would succeed, with exactly the same approach as DARPA. I called this the “renewed Kaizen” process. It is the same idea as Kaizen, but it includes a willingness to fail.

While DARPA has achieved incredible results, most of the time does it fail. There are many lessons learned from the DARPA playbook. High risk is necessary for great success, therefore choosing high-risk, high-reward projects is the first lesson from DARPA. Avoiding choosing projects that duplicate other people's work and focusing on a few well-funded projects versus many smaller projects are also important goals. Additionally, clear quantitative metrics are necessary to measure progress against objectives to ensure that projects are not merely public relations efforts but have potential for real results.

Something unique to DARPA is the high level of trust that the organization puts in its program managers. Each individual program manager at DARPA has the authority to decide how to spend approximately $30 million per year. Another key feature is to establish a culture where reporting honestly is more valued than reporting only good news. Nine out of ten projects are expected to fail, so if anyone is not reporting roughly 90% failure, they are either not telling the truth or choosing projects that are too easy. DARPA is highly selective in choosing staff and there is a great system of rotating program managers, who come from industry or academia, every two to five years, returning to their previous institutions when completing their term.

How can these lessons be applied to Toyota? First, Toyota should become the best company in intelligent vehicles. Second, Toyota’s strengths in design, manufacturing, and user interface technology should be leveraged in order to diversify beyond cars. Also, aiming to become a more vertical company similar to Apple, which went from hardware to software to data, and, recently, to services, is another recommendation. Vertical companies can be very efficient and powerful. Finally, with regard to talent, communicating the attractiveness of working for Toyota is essential to secure highly skilled personnel. Acquisition of talent is actually the most important strength that a company has.

Basically, what I pitched to Toyota seven years ago was that it was important that Toyota established the equivalent of Skunk Works, a famous organization in the U.S., which was part of Lockheed Corporation and later Lockheed Martin, which developed a large amount of innovation in the aerospace business. In my opinion, it was essential for Toyota to make similar disruptive, innovative leaps to deal with the exponential challenges of the future market in the car industry.

Comments and Q＆A

HIRAI Hirohide:
In light of economic security concerns rising worldwide, one topic that I would like to discuss is the role of DARPA in the past and future. How would you describe the current situation with TRI? You also mentioned that the big tech companies are now dominating the innovative arena in the U.S. and attracting talent from around the world. What kind of actions is DARPA taking in this context to attract talent, as you mentioned that it is really the most critical issue for the future success of enterprises?

Gill PRATT:
Acquiring the right talent is very important, but for research organizations—or actually any kind of creative organization—having a few really talented people is much more important than having a large number of people. At TRI we have been successfully attracting people who love Japan and the culture of a company like Toyota, which truly cares about social issues and people. This is an incredible hidden strength that Japan has. The people we have attracted choose to work to us because our mission, culture and country are more attractive to them. Of course, wages must be commensurate with those other companies you mentioned, but we actually don’t need that many of these excellent people.

HIRAI Hirohide:
DARPA has played a great role in nurturing innovation and many of the developed technologies have led to commercial benefits in the U.S. society. From the perspective of industrial policy, what were the strengths of DARPA’s activities? What were some challenges and what does the future of DARPA look like?

Gill PRATT:
Looking back, DARPA’s strength was understanding risk and the need to accept failure. Part of the challenge in any government-funded project is that in case of failure it becomes very political. However, DARPA's charter specifically incorporates taking on high-risk, high-reward projects. By being very honest when reports about successes and failures were made by the DARPA leaders to Congress, the government also learns that failure is part of the process. You have to plant many seeds for the strongest tree to grow, but most of the seeds will actually not result in a good outcome. Politicians usually only want to hear good news, so the constant education of the government and transparency with the public is essential. In terms of the failures that have occurred, there has been failure in transferring the approach of DARPA to other agencies within the U.S. Some have been moderately successful, such as ARPA-E in the energy field, but many smaller efforts to copy DARPA have failed. It is important to understand that half-measures are useless in this domain. One example is the rotation of the personnel. As I mentioned, personnel can stay at DARPA for only between two to five years. Everyone is aware that their time at DARPA is limited, and it causes them to move forward with ideas quickly. This is a vital feature that many of the other efforts did not adopt, and that is one reason that those efforts resulted in failure. Just like with the Toyota production system, many companies tried to copy the system, but ignored essential aspects of it due to a lack of understanding and failed. It is important to understand the entire concept and not simply copy parts of it.

HIRAI Hirohide:
In Japanese culture there is a tendency to be cautious in order to avoid failure and criticism. How do considerations of criticism impact the way innovative high-risk projects are chosen and what are the systems used to make sure that projects are not simply foolish endeavors?

Gill PRATT:
At TRI we call this “opportunity discovery.” How do you find the opportunities where technology may make a big difference? At DARPA, the program managers do not need approval for each individual scientist, company, or school that works for them. Instead, they pitch programs that will have many parties working in competition with each other to find solutions to the problem. Programs are pitched to the director and deputy director of DARPA and reports on the progress are made every year. Sometimes programs are stopped in the middle if they seem feasible. To decide which projects to choose, DARPA takes an approach that is comparable to the Japanese phrase “Genchi Genbutsu”—looking at the situation onsite in other parts of the Defense Department and learning about the greatest challenges they are faced with. Unmanned Aerial Vehicles (UAVs) are one example where DARPA’s investment helped to get the invention over the hump. There is a bottom-up system of trust where ideas are pitched ultimately to the top for approval. The necessity to accept the possibility of failure comes with high-risk, high-reward endeavors and is an integral part of the culture.

HIRAI Hirohide:
When choosing projects, is there awareness that you are competing with China, or Chinese companies, and does this impact the decision-making in terms of which projects are pursued?

Gill PRATT:
Of course, the global competition is looked at. The GPS system is a wonderful example. Many countries very quickly developed jammers that could stop GPS from operating. DARPA had to figure out solutions in response to this technological challenge from adversary nations. The program managers at DARPA are the ones who come up with ideas of how to solve these problems, and most of them are actually young. Unlike many Japanese companies, DARPA has realized that most truly innovative ideas come from young people. However, young people do not always have good judgment, so the role of the more senior people with more experience is to make determinations based on experience, but the creative aspect is handled by young people. This duality of roles is wonderful and can greatly contribute to achieving innovation.

HIRAI Hirohide:
The U.S. government is starting to focus on their industrial policy, especially regarding semiconductors, even though the conventional approach used to be to leave those duties to private companies. What do you think about this current trend?

Gill PRATT:
It has a lot to do with the current international environment. Back in the day when Japan was challenging the U.S. in the semiconductor industry, there actually was a strong government policy action by the U.S. called SEMATECH, a government funded collaboration to develop more sophisticated semiconductors. This was implemented because the U.S. felt threatened and challenged by Japan at the time. It is a cyclic occurrence and now the concern is mostly about China. As a result, diversification and on-shoring of semiconductor facilities are current issues in the U.S. and Japan. Policy is necessary because the industry on its own would not evolve in this direction. Without government intervention, as a result of competition the capitalist system will keep striving for higher efficiency by reducing costs, for example by moving manufacturing to the least expensive areas. However, such an over-optimized solution is actually very fragile. This fragility became noticeable when the COVID-19 pandemic hit. Most of the manufacturing capability of personal protective equipment, such as masks, was located in China, which led to vulnerability when the procurement of protective equipment became increasingly difficult during the pandemic. Governments must recognize that there is always a tradeoff between efficiency and robustness. This is where governments can step in to ensure a diversity of options and a distribution of economic capabilities around the globe. DARPA played a role in this regard in the past with SEMATECH.

HIRAI Hirohide:
You pointed out the importance of attracting and retaining the most talented people. Talent from China has been strengthening Silicon Valley, but—given the current geopolitical situation—is it getting difficult to secure talented people from China? Are there ways of getting similarly talented people from other areas?

Gill PRATT:
Indeed, the strength of Silicon Valley and the reason for its success is its diversity of talented personnel. In general, U.S. immigration policy is extremely open to highly skilled people. There are special visa types for highly skilled individuals. Some come as university students and stay afterwards. The U.S. still has an open door to students from China, so the situation is not bad. It is important for every country to realize what the lesson of Silicon Valley is. Smart and talented people are roughly evenly distributed around the world. The question is, how can we collect the best people? You cannot force people to move, but you can create an attractive living and working environment. It is not only about money, but about opportunity and freedom of thought. Diversity in the geographic background and gender, and other factors are what makes Silicon Valley so strong. Japan, too, has many positive things to offer and my advice would be to create an environment that attracts and welcomes people from all over the world who are at the top of their field.

HIRAI Hirohide:
Silicon Valley companies, and especially start-up companies, are getting increasingly important in the U.S. As the power of start-ups is increasing, what role do they play in terms of industrial policy and carrying ideas developed at DARPA or at colleges into the business world?

Gill PRATT:
Startups are wonderful places, but most of the startups will fail. It is a very similar mindset as in DARPA’s projects. Trying again and again after each failure until you finally succeed is the established culture in Silicon Valley. This is something that the government must be aware of when funding startups. At Toyota, we have a venture fund called Toyota Ventures that is set up in Silicon Valley to fund early-stage startups worldwide. Optimizing the support for startups into established stages to avoid failure and also to ensure that the companies are motivated to succeed is a very important formula. For METI and other Japanese government agencies it is very important to not only study established companies like Google, Facebook, Apple, and Amazon, but also the venture capital startups. Actually, during my time at DARPA, the director was a former venture capitalist, and she used a lot of that experience in funding projects with the same kind of mindset.

FUKUOKA Noriyoshi:
Regarding the relationship between economic security policy by the government and innovation policy of private companies, how can a constructive win-win relation be promoted?

Gill PRATT:
I think the most important point is to realize what the other side cannot do. At DARPA, as a program manager I always had to answer the question when we were considering starting a program: “Why DARPA?” Governments should always ask those questions. What if we do not invest? Will the private industry do it? In the medical field, tropical diseases are a good example. Those diseases mostly occur in countries that are among the poorest in the world so there is not much profit to be made. However, for the U.S. Department of Defense, it is an important topic because they have to operate in all parts of the world, including countries where there are tropical diseases. In the same way, I think it is important for the private industry to also ask the same question and look for ways to be complementary to government approaches, using their specific strengths as contributors to the effort. The private industry should work on projects that only they can do, and the government should fund innovations that the private industry is not going to work on. This type of clear thinking and the symbiotic role between the two is crucial. Of course, there are also projects where everyone needs to cooperate. In such a case, the government can be a catalyst to help in amplifying the private industry’s effort.