Restoring Japan's Leadership In Innovation, Part I
Immobility of Star Scientists Hampers Japan
Source: Hiroshi Shimizu General Purpose Technology, Spin-Out, and Innovation
If I asked people who invented Pfizer’s Covid vaccine, many would reply, “Pfizer, of course.” Some would know it was actually Pfizer’s partner, a German startup called BioNTech, and few would know that BioNTech was founded by two immigrants from Turkey. This is a tale not only of globalization but of a pivotal trend in technological advancement called “open innovation,” a trend where Japan is missing out.
Whereas closed innovation refers to a company trying to invent everything in-house and often rejecting ideas that are “not invented here,” open innovation refers to close mutual collaboration among companies. Sometimes, this entails cooperation among corporate giants, but just as often it means collaboration between giants and startups. It is particularly important in industries undergoing rapid technological change, industries in need of people who can think outside the box. Older giant companies in any country often find it hard to assimilate, or even understand, new ideas that don’t fit into their established business model. So, giants in the US and Europe increasingly see the need to partner with startups.
Half of innovative new drugs developed by America’s pharmaceutical giants these days are invented for them by newer biotech firms, or by a university-biotech partnership. It’s a symbiotic relationship. Since these smaller firms lack the resources to pay for testing, as well as the ability to fabricate and market the drugs, they need to partner with the giants. Among 149 technologies that are subsectors of the information and communications technology (ICT) sector, the average number of patenting firms per technology was a stunning 219. The share of the top three inventing companies was low compared to industries where technological change is not as rapid. So, collaboration is unavoidable if one wants to succeed.
The trend has spread to more prosaic sectors as well. Household goods producer Procter & Gamble relies heavily on open innovation for product development.
Cheaper computers and software have fed this institutional change by reducing the costs of innovation. Back in 1981, 71% of business R&D in the US was conducted by giants with at least 25,000 employees, By 2014 their share had halved to 36% in 2014. Conversely, the share carried out by firms with less than 100 employees had risen to a stunning 8%. Similar changes have occurred in Europe.
Source: Henry Chesbrough Open Innovation: Researching a New Paradigm
This transition has not taken place in Japan. In 2015, only 7% of Japan’s business R&D was conducted by firms with less than 500 employees, compared to 17% in the US, 26% in Korea, and 33% in both France and the UK.
Source: Source: OECD at Business enterprise R-D expenditure by size class and by source of funds https://stats.oecd.org/Index.aspx?DataSetCode=BERD_SIZE
One hurdle in Japan is that the government sends 92% of its financial aid to R&D to companies with more than 500 employees. That’s the least support for startups in the OECD.
Source: OECD (2014) “Figure 5.7. Government-Financed R&D In The Business Sector By Enterprise Size” at http://dx.doi.org/10.1787/888933064601
Japan is Missing Out
It is hard to imagine today’s product lineup in electronics or pharmaceuticals without open innovation. And yet, with a few important exceptions, most leading Japanese companies are missing out on this trend. In Japan, 70% of all R&D is done in-house, and only 0.7% is done in partnership with young firms. The consequences are manifest. No longer is electronics a star industry for Japan because it’s no longer turning out new, must-have products that can compete globally. Despite a 40% surge in global electronics sales from 2008 to 2021, every one of Japan’s top ten electronics hardware manufacturers saw its global sales slump, while total sales of Japanese electronics firms plunged about 30%.
Japan’s problem lies not in a fallback in technological prowess, but in commercial imagination. Producing patents is not the same as producing successful products based on those patents. Japanese companies spend a lot on digital technologies but do not get much bang for all those bucks (or yen). In fact, out of 63 countries in an IMD survey of digital competitiveness, Japan ranked dead last in “business agility,” the ability to benefit commercially from this investment.
Not only are most of Japan’s giants resistant to open innovation but Japan has too few innovative startups to work with. One reason is that new companies cannot get the external financing they need to get off the ground and do the necessary R&D. But, another big reason, is that star scientists and engineers at the bigger companies do not leave to start their own firms, as they do elsewhere. (For more on open innovation and Japan, see this article.)
Immobile Star Scientists
Did you ever wonder why Silicon Valley is in California rather than, say, Microsoft’s home state of Washington? One reason is that California has outlawed non-compete contracts. Such contracts temporarily forbid workers from transferring to a competing firm, or from starting their own firm in competition with their former employer. Washington state, by contrast, enforces non-compete contracts.
Silicon Valley got its start in 1957 when eight star scientists left Shockley Semiconductor Laboratory to create their own new firm, called Fairchild Semiconductor. One of the eight, Robert Noyce, went on to invent the integrated circuit, the foundation of the Personal Computer, at Fairchild in 1960. Then, in 1968, Noyce and Gordon Moore left Fairchild to create Intel, while others founded AMD, LSI Logic, National Semiconductor, and dozens of others. These spinoffs from Fairchild were nicknamed Fairchildren. By 2014, there were 92 of them on the stock market with a combined value of more than $2 trillion. Two of the first venture capital companies were Fairchildren. Within a few years, Santa Clara Valley became known as Silicon Valley.
Japan’s lifetime employment system for regular workers acts like one big non-compete clause, especially at the largest companies. A star scientist or engineer leaving Toyota cannot expect to be hired by Honda, nor even by non-competing top manufacturers. Japan’s star scientists mostly work for very big firms their immobility is even more severe than among Japanese workers in general.
Prof. Hiroshi Shimizu looked at mobility among top scientists and inventors in the laser diode industry, whose most well-known uses are in CDs, DVDs, and optical fibers. The top scientists are measured by those whose patents and papers are most often cited by others. Among the top laser diode 50 scientists in each country, only five of the Japanese transferred to another company. By contrast, 26 of the Americans transferred at least once and some did two or three times. In another list of 83 top Japanese laser diode inventors, only one left his employer to move to a startup. By contrast, among 134 top American inventors, 108 moved to startups. Among all industries, the proportion of R&D specialists who had changed jobs in the 1990s was 43% in Germany, 38% in the U.S., and 35% in the UK, but a mere 6% in Japan (see chart at the top).
In addition to the usual disincentives for labor mobility in Japan, star scientists face another hurdle. A great deal of their knowledge is implicit learning as a member of a team and may even be company-specific. Unless they leave together with others in the team, they lose that part of their human capital.
Shimizu compares the relationship between incumbents and spinouts to a fruit-bearing tree. A tree needs a strong trunk to develop a general technology that can have a myriad of applications, whether it be a steam engine or a laser diode. The applications are its fruit. In most rich countries these days, big incumbents work on the trunk while spinouts develop the fruit. If too many top personnel leave the big incumbents too soon, the growth of the trunk will be stunted. If none leave, the variety and abundance of fruit will be constricted. In the laser diode sector, Japan’s big incumbents controlled both trunk and fruit, often doing so on the same technological trajectory with other incumbents. When that trajectory no longer suited a changing market and prices dropped, few could adapt, and they exited the sector altogether. What seemed rational to each firm, and its star scientists, did not serve them in the long run.
One Route to Change
How might this change? One possible initial driver is the growth of university spinouts. In 1999, Tokyo changed Japanese law to emulate America’s 1980 Bayh-Dole Act, which allows universities to own and commercialize research done by their faculty in order to foster spinoff companies. The number of such spinoffs in Japan is small but growing. As of 2019, there were more than 2,500, of which relatively few had been listed on the stock market as yet.
Most of these spinoffs are still tiny, with half having no more than five employees and most of the rest between 5 and 20 staffers. They’re the ones who need government financial aid for R&D, but do not get it. At half of these university-initiated startups, the CEO as well as the Chief Technology Officer have a background in academic research in the natural sciences. Importantly, another 18% of CEOs are former managers at larger companies and an additional 24% had been managers at SMEs. That’s encouraging. While scientists may have a vision of the product and its potential, few have the business acumen to turn a great technological vision into a viable commercial enterprise.
If this process reaches a critical mass, it could lead to star scientists at big private companies following suit.
One work-around: doing your R&D outside Japan. Feeding into that need is the shrinking number of college graduates to staff R&D inside Japan. From firm visits as the judge for an innovation competition, I know several examples in automotive for Japanese suppliers doing core work in the US and Europe. Of course this may be successful but not generate value-added for the domestic Japanese economy. Inside Japan, well, mobility is indeed low, though "shukko" and other work-arounds may mean it is _slightly_ better than your data suggestion.
FYI Ministry of Economy, Trade and Industry (Japan), “IT-jinzai jukyu ni kan suru chosa (gaiyo)” [Study on the Supply and Demand of IT Personnel (Summary)], October 20, 2022, https://www.meti.go.jp/policy/it_policy/jinzai/gaiyou.pdf.