How Does The Dearth of PhDs Hinder Japanese Corporate Innovation

Also Hindered by Immobility of Scientists, Especially To Startups

Source: Hiroshi Shimizu General Purpose Technology, Spin-Out, and Innovation

What does it matter if Japanese companies don’t hire PhDs? Some readers asked this in reaction to a recent post. I gave a brief answer in this addendum that showed the superior innovative capabilities of Japanese firms that hire PhDs. In this post, I’d like to probe this issue more deeply. I’ll also discuss how innovation is hindered by how few star scientists move from firm to firm, especially not to startups.

A World Of Science-Based Technology

We’re no longer living in the world of Thomas Edison and the Wright brothers, a world in which mechanical geniuses without any background in science could come up with breakthrough innovations, from making the first light bulb that would not burn out in minutes to getting an airplane off the ground. Yet, Edison worked at a time when technology was increasingly reliant on scientific discoveries, and he knew it. For example, neither the Bessemer process in steel (1856) nor the open-hearth furnace (1861) could have been created without the sciences of chemistry and metallurgy, which allowed steelmakers to remove impurities. Moreover, even this steel could not be widely used without a host of additional science-based improvements. Technology historian Nathan Rosenberg wrote:

“The quality of the [steel] was highly sensitive to even minute variations in the composition of the inputs [such as sulfur, phosphorus, and even nitrogen from the air—rk]... In anticipation of the problems,...a chemical laboratory was established [at the first Bessemer plant in 1863]. This was... one of the first laboratories attached to any industrial firm. Similarly, large users of Bessemer steel were likely to set up testing laboratories ... Such was the case of the railroads [to ensure that steel tracks could tolerate frequent use with heavier loads than in the past].”

Recognizing this transition, Edison himself employed university-trained scientists in his second big R&D lab (1887). Today, technology is more reliant on science than ever (and vice versa). Without quantum physics, we would not have the microchips that drive the modern economy; without understanding DNA, no Covid vaccine; without chemistry and genetics, none of the modern fertilizers that enable food output to multiply much faster than population. At the same time, without today’s technologies, much of today’s scientific achievements would be impossible.

PhDs in American Commercial Innovation

Given all this, it should not surprise us that half of America’s inventors these days are PhDs in science or engineering. Nor that 35 of America’s billionaires obtained PhDs before going into business, many of them as company founders. Nor that a surprising 16% of Google employees have PhDs, along with 11% at Facebook and 10% at Uber. When we expand our view to the 650 US-born CEOs and heads of product development in 500 engineering and technology companies, we find that 10% hold PhDs. Among 347 American unicorns, 18% had at least one PhD co-founder.

One might think that all these PhDs are in STEM (Science, Technology, Engineering and Math) fields. Most are, but a goodly number are in the humanities, whose expertise is in people, i.e. customers and staffers. It often takes a combination of these two mindsets to yield commercial success. Over a decade ago, Google bought an artificial intelligence (AI) company called Aardvark, whose founder (Damon Horowitz) could not master AI until he earned his PhD in philosophy. Google hired him, making him “In-House Philosopher / Director of Engineering.”

Dearth Of PhDs In Japanese Companies

It’s hard to get data on how many Japanese business leaders have a PhD, so I am likely missing several. But so far, I’ve found only a few “unicorns” or other leading startups whose founders include a PhD, such as Smart News, Cyberdyne, Euglena, and Spiber. So far, I’ve not found a single PhD among the CEOs at leading traditional companies. Someone can correct me if they know of any.

Now, let’s look at Japan’s inventors. In Japan, just 10% hold PhDs, compared to half in America. In a survey of 1,100 large Japanese firms, only 186 employed any PhDs. Of the 23,538 new researchers recruited by Japanese companies in 2016, only 904 (4%) held PhDs.

Do University-Trained PhDs Make A Difference?

(This section repeats material from my December “addendum.”)

Hiring PhDs makes a big difference. According to a survey by Professors Yuya Ikeda and Tomohiko Inui, among large companies, 40% of those employing at least some PhDs had created significant product innovations, and 43% had created process innovations. That’s more than double the 20% rate at firms with no PhDs. There were similar, albeit a bit smaller, results among mid-sized firms.

When one considers other factors in innovation—R&D budget, firm size, selling in foreign markets, etc.—the additional benefit of employing PhDs added 9 to 14 percentage points (among large and medium firms, respectively) to the share of firms developing important product innovations and 14 to 12 points in process innovations. So, having PhDs accounted for around half of the difference in company innovation.

In Japan, a large share of the PhDs employed at companies did not get their doctorates at a university but by doing a dissertation based on their research at their employer. They were initially hired at the bachelor's or master’s level. These “dissertation only (DO)” PhDs account for half of all PhD inventors in Japan. However, a study by Koichiro Onishi and Sadao Nagaoka found that these DO PhDs were substantially less creative and productive than university-trained PhDs. Not only did they develop fewer patents, but their patents and papers were cited less often by other inventors.

Company Absorption of Others’ Knowledge

No technologist creates innovations de novo. All stand on the shoulders of their predecessors and peers. So, the ability of a company to absorb external knowledge is indispensable to further advances. As former METI veteran Koichi Akaishi explained to me (see this post), not hiring PhDs means that company researchers often don’t understand university scientists well enough to figure out how to use that science to create commercial value. He pointed to the example of quantum computing.

In a world where knowledge created anywhere is available everywhere, the ability to absorb external knowledge and see its commercial potential is perhaps even more critical to a country’s economic growth than doing basic research. Akira Morita of SONY did not invent the transistor, but he saw how to turn silicon into commercial gold. Onishi and Nagaoka state that university-trained PhDs “are more likely [than other technologists, including DOs] to generate internal knowledge spillover within a company through absorbing the external scientific knowledge.”

Let’s compare Japan’s situation to the so-called European Paradox, the notion that, “while EU scientific performance is on par with that of its main international competitors, Europe lags behind in converting research results into [commercial] innovations and gaining competitive advantage.” The authors of this study looked at the research produced by the European Research Council (ERC), a funding agency created by the EU in 2007 to finance frontier research. Its 10,000 grants have led to 200,000 articles in scientific journals, a dozen Nobel prizes, and more inventions than other European science.

But who took the most economic advantage of all this knowledge? Not European companies, but American ones, particularly startups. “In absolute terms, however, the corporate sector, especially US companies, still leads in converting these research findings into valuable inventions [emphasis added—rk]...Thus, although we document a robust spillover effect of ERC science on inventive activities, European industry still lags behind in fully capitalizing on [this] spillover.”

Even considering the deterioration in Japan’s basic research, I'd argue that Japan’s biggest technological problem is an analogous “Japanese Paradox.”

The Critical Role Of Startups; Immobility of Japan’s Star Scientists

Digging a little deeper, the authors of the ERC study looked at the reasons American companies gained more than Europeans. “The difference...is primarily driven by startup patents, with the US taking the lead.” These American startups, they state, “spearhead invention activities at the forefront of science.”

By contrast, Japan’s dearth of science-based startups is a big contributor to the decline in the country’s commercial innovations.

This raises yet another issue: the difficulty Japanese startups have in either being founded by its brainiest scientists or in being able to hire star PhDs. The rigidity of lifetime employment is even greater in science than in general. Not only do few scientists move from firm to firm or from university to firm, but of the few who do move, hardly any shift to startups.

The chart at the top of the post compares the top 1% of scientists working on laser diodes, 100 in the US and 90 in Japan. These diodes are pivotal to a myriad of products. Only 26 of the Americans stayed with their original employer for their entire career. Eighty-two of the Japanese did so, most of whom worked for big traditional firms. Of the 8 Japanese who did move, only one moved to a startup. Forty-one of the Americans moved from an incumbent firm to a startup, while 52 moved from one startup to another, some more than once, and 15 moved from a university to a startup. As I discussed in a previous post, the immobility of star scientists is typical throughout Japanese business.

This raises another issue. The PhD problem can affect what kind of startups Japan generates. We know that the difficulty in getting external finance leads Japanese startups to avoid ventures that require hundreds of millions, or billions, of dollars in startup money (as Tesla needed). Do Japan’s startups also avoid projects needing university-trained PhDs, the kind of ventures that produce breakthrough products and companies with sales in the billions of dollars?

How The Mobility of Scientists Promotes Growth

Countries and regions where scientists are more mobile tend to innovate more and thus grow faster as knowledge of varied ideas spreads from firm to firm. Have you ever wondered why Silicon Valley sprouted in California? Partly, it’s because California bans non-compete clauses in labor contracts. These prevent employees who leave a firm from working in the same field for a few years.

Why, then, do traditional firms worldwide, including Japan, try to prevent their R&D workers from leaving, some using non-compete clauses? That’s because they believe their departure will hurt their company while benefiting their competitors. If they are correct, then what is good for the country as a whole is not good for its leading firms. But are they right?

A 2014 study shows that this widely-held belief is not necessarily true. In Denmark, “labor mobility increases the total innovative activity of [both] the new and the old employer.” While the biggest benefits went to the new firm—because the new R&D worker brought fresh ideas—the firm that the technologist left also saw a significant increase in the number of patents. One reason is that the firm from which the R&D left paid greater attention to the research of the firm that the worker went to, which led to both the old and the new employer citing each other in subsequent patents. The departing workers contributed more to patenting than the workers who stayed because each firm got fresh ideas from the interaction between the two firms and their R&D workers.

Of course, all of this positive feedback requires that the firms involved be willing to engage in “open innovation”—i.e., deep technical collaboration with other firms—and that is something from which most Japanese traditional firms still shy away.

To provide more support, donate several subs at $50 each. You do not need to name the sub recipients, just how many. This is a one-time contribution; it does not repeat automatically. Please click the button below.