Net Assessing American and Chinese Innovation
Simon Leys, the Belgian sinologist, polymath, and translator of Confucius’ Analects, once noted that “Western incapacity to grasp the Soviet reality and all its Asian variants was not a failure of information (which was always plentiful); it was a failure of imagination.” Leys referred to the moral obtuseness of much of the West’s intelligentsia, its failure to grasp the scale and iniquity of the catastrophes inflicted by Communist regimes.
But there are varieties of such bafflement. Presently, there is a failure to conceive of the scope of U.S. or Chinese capabilities — we simply do not know what either America or China can and cannot do. We especially have not taken a proper accounting of the two countries’ comparative capabilities in innovation. Such innovation is essential to global economic leadership, and that is what China, whose rise has been markedly reliant on economic growth, aspires to.
As of this date, the U.S. government has yet to appropriately examine — using net assessment, the most comprehensive method — the U.S. and Chinese innovation ecosystems. No assessment exists either of the ecosystems in their totalities, of the various types of innovation, or of any particular industrial and technological sectors within the ecosystems. And because our current understandings are muddled, our future imaginings are limited.
Such failures to grasp a competitor’s relative strengths and weaknesses have occurred before. Throughout the 1960s and into the 1970s, CIA analysts continually overestimated Soviet economic power in relation to the United States. One of the most pervasive misunderstandings was the size of the Soviet military in proportion to the size of the overall economy. Common wisdom was that the burden of defense upon the Soviet economy was roughly comparable to that in the United States — from 6 to 8 percent.
Looking at Soviet military power tank for tank or missile for missile, it looked formidable indeed. But these “apples to apples” threat assessments did not tell the real story. Other analysts who worked for the Department of Defense’s Office of Net Assessment provided more detailed, contextualized readings that showed that the Soviet economic system was distorted and failing. Military spending was anywhere from 15 to 50 percent of overall GDP. The juggernaut that looked so formidable was built on an economic house of cards.
How did the analysts’ assessments come to such different conclusions? Net assessment is very different from “threat” assessment. It is not focused on threats and attendant reactions. Threat assessments produce crisis responses — and, too often, attendant crisis psychology.
Generally net assessment is empirically based and data-driven, interdisciplinary, and diagnostic rather than prescriptive. This last trait is especially notable. Net assessments are not supposed to provide laundry lists of proposed policy solutions. Such “solutioning,” to paraphrase Andrew Marshall, net assessment’s founding father, corrupts the analysis. The quest for solution distorts information and diminishes real understanding.
Net assessment’s origins are typically traced to the early 1970s, with the arrival of Marshall at the Pentagon at the behest of then-Secretary of Defense James Schlesinger. Marshall had begun his career as a Rand analyst, and while there began to look beyond military circles for new ways of thinking about national security issues. He gained insights from business culture, especially from Harvard Business School’s case study method, which embedded business decisions in a variety of practices and behaviors that went well beyond typical economic “rational choice” modeling.
Marshall’s notion about diagnosing and not prescribing appears especially salient regarding China, a great power that has seemingly defied so many expectations and understandings for a long time. There have been recent “strategic net assessments” that have focused on China. A recent one provided a whole list of recommendations — somewhat at odds with Marshall’s anti-prescriptive edict. And it seemed overwhelmingly focused on pure geopolitics, with primarily diplomatic and alternative “military political approaches.”
But any net assessment of China, a country that has had in the post-Mao era historically unprecedented economic growth, cannot be only an exercise in geopolitical or militarized reductionism. This is not to say that China’s military capacity is unimportant. However, Philip Saunders contends that Chinese military reform is not focused on winning an arms race per se with the United States. There are other imperatives driving China’s military reforms: regional security needs to be maintained, interservice rivalries need to be mitigated, and Belt and Road lines of communication need to be protected.
Rather, it appears highly plausible that what is driving China as a nation-state, both internally and toward the world at large, is its effort to become the world’s leading economic superpower, using innovation as the source of its growth and security. Such innovation is the basis for what Tai Ming Cheung terms China’s “techno-security state.” This involves a different state-market relationship than the one that currently exists in China today, an overhaul of the human capital management system, a more robust policy-legal enforcement system, improved international science and technology cooperation, and the degradation of competitors’ technology firms.
The Picture Doesn’t Look Good
What is most pertinent, then, is a comprehensive net assessment of both the U.S. and Chinese innovation ecosystems. Innovation is distinct from simply “invention.” Whereas the latter is merely a good idea, the former is something brought to fruition for the market and put to public use. It can be a new product, such as the smartphone, or it can be a new process, such as Toyota’s transformation of automobile assembly. Regardless, innovation requires a system to make it happen. A national innovation system is that national network of institutions and relationships that fosters the innovation process, even better described as an ecosystem. It is a dynamic, interconnected system occurring at both the national and regional level — sometimes termed the “triple helix” — in which government, industry, and academic institutions all influence and interact with each other.
There certainly appears to be a systematic effort within the Chinese Communist Party and the Xi regime toward strengthening China’s innovation ecosystem. Whether Chinese centers in Shanghai or the Pearl River Delta can truly rival Silicon Valley or Boston’s Route 128 technology corridor is likely dubious, at least for now. But it appears that China has mounted a multi-sector effort at achieving original, product-focused innovation in an array of industries, from shipping to quantum computing to agribusiness.
Correspondingly, if the Chinese innovation ecosystem is apparently strengthening, the American ecosystem looks to be moving in the opposite direction. The innovation ecosystem championed by the likes of Vannevar Bush at its Cold War peak shows decline. Military-driven innovation has fallen off, as defense industries have moved out of the general economy to become, through vertical and horizontal integration, the so-called defense “primes.” The American venture capitalist system seems unable to commit to innovation on a significant scale — too much committed, in Peter Thiel’s famous phrase, to bits rather than atoms, on incremental improvements such as apps in smartphones and not “big” innovation in things like energy and space travel, whether due to high hurdle rates (the minimum rates of returns that firms require for investment), the so-called valley of death (the lengthy time required to profit from significant investment), or some other financial inhibition.
The notions of American “declinism” and weakness relative to China that the above reading suggests have of course been hotly disputed. Peter Zeihan insists that China is “overhyped” and rife with problems ranging from low birth rates to inadequate energy access. Michael Beckley argues that Americans are, overall, many times more productive, healthier, and better educated. The point, however, is that understandings of the Chinese and American innovation ecosystems remain incomplete hypotheses that need to be fully interrogated by analysts who possess a deep understanding of firms, industries, technologies, and technology systems, in a net assessment that takes into account real and potential human capital, and that takes the long view to see what is historically trending and what is merely transient.
How to Fill in the Details
There is no absolute standard or model blueprint for conducting a net assessment of innovation potential. Perhaps the most comprehensive approach is found in the Organisation for Economic Cooperation and Development’s Oslo Manual that measures a country’s capacity to innovate. It provides a series of categories to subdivide innovation: 1) by innovation sector; 2) by innovation type (a product, process, market, or organizational type); and 3) by region, firm size, and other less holistic (though still significant) categories. You could use this scale to assess a single national industry’s innovation ecosystem — such as aviation or robotics — with the types delineated, and categorized by firm size, as a start. This would be in keeping with net assessment as very much an iterative, evolving practice, with any single assessment viewed as simply one tile in a mosaic that accumulates meaning and understanding with additional data.
First: Examine Behaviors
To be appropriately empirical and multidisciplinary, the first things that analysts should evaluate are the respective innovation ecosystems behaviors — what entities within the system are actually doing. What sort of behaviors should be examined? The first data to be evaluated are the system’s inputs. Robert Atkinson lists four key metrics: overall research and development investment; so-called research and development “intensity” (that is, investment as a percentage of GDP); business research and development investment; and government research and development investment (separately categorized as defense and non-defense government investment). Analysts should look back at least a decade to find trends and patterns and to provide deeper context. The assessment should also determine the actual number of firms in the United States and China that are investing in high-tech innovation such as AI, along with their respective sizes, as well as such firms’ actual particular research and development investments. U.S. and Chinese firms could be rank-ordered to see where the two nations’ various entities stand up in terms of research and development investment globally.
These inputs should be complemented by output data that indicates “results” in the respective ecosystems. Atkinson considers many of the relevant results to be human capital information —researchers per capita, annual numbers of science, technology, engineering, and math (STEM) graduates, and top-ranked science and technology-oriented universities. Output data also includes the annual number of high-tech patents and published high-tech papers, the annual amount of high-tech exports, and the annual number of new-start companies. Likewise, this output data should not simply be analyzed as a snapshot in time, but placed in historical context.
Both input and output data lead to higher-order questions and, hopefully, to more meaningful behavioral data. It is not only about determining, for example, the number of patents granted, but also whether patent licenses are effectively transferred to industry for commercialization. This is the kind of outcome data that determines whether behaviors in the innovation ecosystem are effective, which includes their effects on matters such as employment and value-added productivity. It is also the data that is most difficult to obtain. Within the U.S. system, though, there are some promising capabilities. TechLink, for example, a technology intermediary headquartered at Montana State University that works with government and the private sector, recently evaluated the effectiveness of Department of Defense patent license agreements transferred to private industry and was able to determine outcome data such as the jobs created, tax revenues, new product sales, and nationwide economic impact of the patents.
Second: Examine Intentions
Only afterwards can intentions — the various plans, strategies, and regulations — be analyzed, so as not to read intentions into behaviors too soon. What is actually happening and what leaders say is or should be happening are often very different. The intention/behavior gap could be the result of the ecosystem moving faster than the authorities’ ability to define or to control it. For example, the pace of innovation development in China, especially in technological-industrial hubs such as Beijing, has sometimes been so rapid that it has surpassed the ability of Chinese institutional bureaucracy and attendant laws to keep pace.
The best way to unravel the strands is to use Atkinson’s “innovation success triangle” with its three sides that represent the business environment’s various requirements; the national trade, tax, and regulatory schemes; and the national innovation ecosystem’s policies. Business environment requirements foster or inhibit research and development investment, establish requirements for venture capital enterprises (which could include so-called hurdle rates), and set guidance for start-up companies. The trade, tax, and regulatory schemes include tax burdens or credits provided to innovating firms, patent system rules, product regulations, and governmental policies regarding foreign investment. The ecosystem innovation policies include legislation that fosters innovation infrastructure development, and strategies, plans, and policies that articulate national agendas.
With this triangle of innovation intention explored, net assessment requires higher-order synthesis. Beyond simply listing what leaders have said about innovation or what policies have been laid out, what is the contextual narrative created by the web of statements? What do the intentions say about the role of innovation as part of the national economy and the nation’s future? What is the relationship between the intentional and the behavioral data — is what is actually happening in accordance with the expressed intentions?
Third: Consciously Articulate Perceptual Limits
Innovation is continuous and inherently difficult to capture. The creative destruction that innovation wreaks can be massively destabilizing and wildly unpredictable. Comparing nations as different as China and the United States may create major misperceptions. Any such innovation ecosystem assessment should therefore be conscious of perceptual limitations. Analysts should constantly and reflexively acknowledge that any assessment can be creating a false impression that obscures the real story.
That knowledge of the Chinese system is limited is hardly surprising. Chinese growth figures may be significantly inflated. Distinctions between regularized and shadow banks are difficult to discern. The flexibility and transparency of state-owned enterprises are still unclear. Western misperceptions have historically compounded misunderstandings. One common problem throughout the 1990s was to view China using standard market analysis and metrics. Standardized methods would score China as less or slightly more economically “free” than, say, Russia or other former Soviet bloc nations, implying that China was simply one “emerging market” among many, hoping to join the rest of the developed world on terms that the developed world mandated.
And America’s perceptual limits about itself are also very real. The decentralized federalism, openness, and democratic process in the United States mean that there is less consistency and constancy in innovation policies. Much innovation occurs in the private sector and information therein is proprietary. Valuable tools are scattered throughout the system (such as the aforementioned TechLink data surveys), but only so much can be reasonably ascertained. The point is not to despair at the process of assessment, but to be conscious of its limitations. No matter how thorough, no assessment can find everything.
Fourth: Assess Strengths and Weaknesses and Look for “Asymmetries”
Finally, rather than proceeding on to a set of recommendations, proposals, or strategies, the assessment should identify asymmetries between the ecosystems, and refine them through a series of further questions. The goal here is to determine, for example, relative American strength to relative Chinese weakness in the ecosystem. This is somewhat like the business practice of “SWOT” (strength, weakness, opportunity, threat) analysis, though with a significant difference. This is a comparative analysis between competing, complex national ecosystems, and not simply internal firm analysis. And asymmetries can be determined not only from, say, a comparative analysis of the countries’ overall exporting of hi-tech products, but also from those products’ diversification. For example, China exports a lot of high technology. But the vast majority of those exports are information technology hardware, with some estimates suggesting that this makes up more than 90 percent of its high-tech exports. This could indicate a lack of diversification in the hi-tech innovation sector — and a resultant possible significant innovation asymmetry with the United States.
Another example of an asymmetry to explore is what might look at first glance like a potential strength for the United States: its prestigious science and technology-focused universities, and the (at present) comparative weakness of China’s. But the U.S. system is experiencing an apparent decline in the numbers of STEM graduate students while the number of Chinese STEM graduate students is increasing. With such an asymmetrical relationship identified, we can shake out its implications in a series of refining questions.
It is at this point that net assessment ceases, in order not to prescribe and possibly taint the analysis. Whether or not any grander strategies, plans, or policies do finally emerge, an alternative possibility emerges from the net assessment process itself. Instead of some clumsy, unwieldy grand strategy, such asymmetrical identifications could yield more flexible and adjustable stratagems — engagements, refinements, and calibrations within the ecosystem, rather than potentially misguided efforts to control, manage, or master it.
Willing Blindness
At present the United States knows far too little about itself and about China, its chief technological rival, when it comes to innovation. As a whole, policymakers and strategists have not adequately studied and understood the key source of American and Chinese power in today’s world — innovation capacity — precisely because they have not net-assessed that capacity. That first step remains to be taken. Perhaps the approach recommended here provides a path.
Walter M. Hudson is an associate professor at the National Defense University’s Eisenhower School for National Security and Resource Strategy where he has taught courses on strategy, strategic leadership, and geo-economic competition. He is also a global fellow at the Woodrow Wilson International Center for Scholars. The opinions expressed in this article are his personal views and do not reflect those of the Department of Defense or National Defense University.
