Technological Innovation in China: Current State and Challenges

Abstract

Technological progress and the substantial easing of international competition after the Cold War have jointly shaped the pattern of division of labor in the international industry chain over the past few decades. While serial production and network collaboration have improved production efficiency, industry chain security has gradually become a cause for concern. We believe the policy response should focus on enhancing irreplaceability domestically to address risks from international competition as countries deemed irreplaceable in the industry chain could resume production by establishing new network connections with other production partners, while countries that are less irreplaceable are at risk of being marginalized or even removed from the production network.

According to our analysis of the global value chain, China as a whole is in the downstream. In our view, on one hand, China's advantages mainly lie in low-tech manufacturing industries, exposing the country to potential horizontal risks posed by trade frictions for the purpose of decentralization. On the other hand, China's disadvantages mainly lie in primary products such as oil and medium- and high-tech industries such as software, exposing the country to potential vertical risks posed by overseas supply bans.

The key to coping with horizontal and vertical risks lies in technological progress. To clarify the future path of innovation, we believe it is crucial to examine the current state of China’s technological innovation and understand its major challenges. We take the perspective of a country’s innovation system to assess the status quo of China’s innovation activity. Regarding innovation input, China ranks only behind the US in terms of R&D spending, but still lags behind the world’s technology powerhouses in terms of R&D intensity. We think the country still needs to enhance its R&D intensity, especially in basic research that is vital for radical innovation. Talent is another crucial driver of innovation. China has a large talent base, but we believe there is room for improvement in the number of researchers per thousand people in the labor force as well as the quality of labor and structure of human capital. We believe the supply-side challenges constraining the education system and hindering the cultivation of innovators are three-fold: Uneven distribution of high-quality basic education resources, the intensified incentive system in exam-oriented education, and educational administration factors in research-oriented universities.

In terms of technological outputs, we focus on academic publications and commercial patent applications to measure China’s achievements in knowledge innovation. We compare the 30-year history of China's knowledge innovation with international experience and highlight five patterns. First, as a latecomer to knowledge innovation, China has caught up with developed countries in terms of quantity, and the gap between China and developed countries in terms of quality is narrowing. Second, China’s contribution of core innovation to global output remains limited and it lags behind developed countries in terms of originality. Third, from the perspective of conversion of achievements in knowledge innovation, industries do a good job while conversion of academic research results is relatively weak. Fourth, China tends to post strong performance (in knowledge innovation) in areas in which it has comparative advantages in international trade. Fifth, the growth momentum of China’s cooperation with foreign entities in knowledge innovation has slowed in recent years, especially in the field of academic cooperation, which deviates from the global trend.

Having examined China’s current position in the global value chain as well as its innovation inputs and performance, we use criteria of “systemic importance” and “promising development prospects” to identify three major areas for further analysis: The digital economy, the green economy, and the bioeconomy. We note that regardless of international competition, China still needs to increase R&D intensity and accelerate technological progress to achieve its medium- and long-term growth targets. Regardless of the perspective of either international competition or domestic growth, China needs to accelerate technological progress, in our view, and policy intervention is indispensable to accelerating technological innovation.

In this book, we discuss the paths to accelerating technological innovation in two parts. Three macro chapters focus on the analysis of innovation economics including the status quo (this chapter), the scale effect (Chap. 2), and the innovation ecosystem (Chap. 7). We also include four industry-focused chapters, which examine the application of innovation economics in digital industries (Chap. 3), green energy (Chap. 4), the biotech industries (Chap. 5), and supporting industries such as manufacturing and logistics (Chap. 6).

1.1 Innovation in the Era of Heightened International Competition

With technological advancements in manufacturing and logistics, product trade has gradually evolved into industry chain trade. From our perspective, while division of labor in the industry chain has improved efficiency, the increase in segments has made the entire industry chain and the economy more vulnerable. The COVID-19 pandemic has accelerated the exposure of this problem,¹ prompting countries to begin paying close attention to industry chain security. For China, the industry chain risks emerged earlier—trade frictions in 2018 underscored the importance of industry chain security. Regarding how to ensure industry chain security, countries seem to have reached a macro-level consensus that it is necessary to shorten the industry chain and increase local production capacity, gaining security at the cost of a certain loss in efficiency.

According to this macro point of view, efficiency and security cannot be achieved at the same time. However, from a micro perspective, we believe an important basis for the industry chain division of labor is the Wintel model.² This production model breaks up a production process originally within a single company into steps carried out in different companies, regions or even countries, so as to form new divisions of labor and new production systems. Economist Masahiko Aoki, professor emeritus both at Stanford University in the US and at Kyoto University in Japan, studied the US's new economy and concluded that the US relied on new forms of industrial organization represented by the Wintel model to reverse the competitive disadvantages of US companies against Japanese companies, and achieve sustained rapid economic growth in the mid to late 1990s³ (Fig. 1.1).

Fig. 1.1

Evolution of manufacturing production models. Source Wintel model: micro basis of US new economy and global industrial restructuring (Huang Weiping et al. 2004), CICC Global Institute

However, for industries characterized by the Wintel model, the irreplaceability of each segment in the industry chain varies. R&D is at the core of the industry chain and is the most irreplaceable segment. A company can take a dominant position to a great extent along the entire value chain if it dominates upstream R&D. Other segments with lower irreplaceability provide production support for the core segment and are subordinate to it. Judging from the Wintel model at the micro level, the key to improving industry chain security is to increase one's irreplaceability. In our opinion, this means that efficiency and security can be achieved at the same time, and the key lies in strengthening R&D and technological innovation.

The problem is that the above two perspectives derive different views of the relationship between efficiency and security of the industry chain. Hence, we will explore and assess these differences in this chapter.

The essence of industry chain security is high irreplaceability. Theoretically speaking, industry chain division of labor can take place both domestically and internationally, so the security of the industry chain can also be discussed from the perspectives of domestic production and international trade. In reality, concerns about industry chain security mainly arise from the perspective of international trade. In June 2021, the White House stated in its supply chain security review report that concerns about industry chain security stem from a reliance on foreign sources of raw materials and foreign manufacturing production.⁴ Similarly, in China, discussions on this topic have been on the rise, with the number of papers on the topic of industry chain security being also highly related to the volume of China's exports and imports (Fig. 1.2). Therefore, we also discuss industry chain security from the perspective of international trade.

Fig. 1.2

China's domestic discussion on industry chain security. Note CNKI stands for China National Knowledge Infrastructure. Source cnki.net, Wind Info, CICC Global Institute

1.1.1 Characteristics of Global Industry Chain: Serial Production and International Collaboration

The global industry chain has formed a complex production structure characterized by serial production and international collaboration. Comparing recent global trade flows with those 20 years ago, we find various economies are increasingly involved in the global trade division of labor, and the global trade network has become increasingly complex. During this process, China has replaced Japan as the Asian center of the global trade network (Fig. 1.3).

Fig. 1.3

Global trade flows. Note The chart shows bilateral trade of more than US$5bn (measured in 2000 constant US dollars). Any curve represents the export of goods from the upstream node to the downstream node in a clockwise direction. The size of the node represents the total export value of the economy, and the thickness of the curve represents the size of the bilateral trade volume. The color of the curve is the same as the color of the exporting country. Source Gross Trade Accounting: Official Trade Statistics and Measurement of Global Value Chains (Wang Zhi et al. 2015), ADB MRIO database, CICC Global Institute

The characteristics of this industrial division of labor are jointly shaped by technological progress and thawing of international relations. In terms of technological progress, the most important changes are a decline in logistics costs and the adoption of new manufacturing models, notably the Wintel model. The decline in logistics costs makes global resource allocation and cross-border collaborative production possible. The Wintel model makes cross-enterprise serial production a common production model in technology-intensive industries. Under the influence of these two factors, the division of labor of the global industry chain has become increasingly sophisticated, and the global value chain has continued to lengthen.

In terms of international relations, the easing of geopolitical conflicts after World War II (especially after the Cold War) created a favorable international environment for the signing of trade agreements, which greatly lowered tariff levels worldwide and reduced the cost of cross-border flows of goods. The connection of different regional markets has created globally integrated new product and labor markets, promoting the global industrial division of labor. In particular, China's accession to the World Trade Organization (WTO) in 2001 has profoundly changed the global and regional trade landscapes as well as division of labor, and the network-based production featuring collaboration between countries has become increasingly complex.

1.1.2 Sources of Industry Chain Risk and the Nature of Security

As the division of labor becomes increasingly sophisticated, the industry chain also faces various risks. The Biden Administration states that sources of industry chain risk include global pandemics, cyberattacks, climate shocks and extreme weather, terrorist attacks, geopolitical and economic competition, and other factors that will erode the US’s key manufacturing capabilities and impact the availability and integrity of critical goods, products, and services.⁵ In summary, industry chain risks can be divided into two categories.

First, natural disasters impact serial production, and disruption of any segment will make it difficult to complete the entire production process. For example, in March 2011, a magnitude 9.0 earthquake in northeastern Japan caused many wafer fabs to reduce production or even shut down, triggering a global integrated circuit sales price increase of about 8% MoM and a flash memory price increase of nearly 25% MoM.

Second, international competition disrupts network collaboration, forcing a country's production nodes to reconnect with other nodes, and the international collaboration network is artificially impacted so that production efficiency decreases. Upon the introduction of trade barriers, technology blockades, and trade bans on raw materials between countries originally in a collaborative relationship, the original production model of the industry chain could be severely impacted. This risk is particularly prominent against the background of rivalry between major powers, in our opinion.

Compared with natural disasters, we believe risks of international competition deserve more attention, for the following reasons. First, natural disasters usually do not last long. For example, the impact of the Japanese earthquake in March 2011 on semiconductor prices largely subsided after the end of the month. In contrast, major international competition usually lasts longer. Second, under the impact of natural disasters, the duration of the impact and the degree of damage can be minimized through close cooperation between countries. In contrast, international competition alone is enough to disrupt the entire industry chain even if there is no natural disaster.

For industry chain risks caused by international competition, there are no apparent winners in the short term as no one can complete serial production without cooperation from others. However, in the medium and long term, the impact on different segments varies: On the one hand, the irreplaceability of each segment in the industry chain is different. On the other hand, the division of labor in the industry chain is characterized both by serial production and by network collaboration. Countries in highly irreplaceable segments are capable of resuming production by gradually establishing new connections with other production nodes. However, countries in less irreplaceable segments may be gradually marginalized by the industry chain, and may even be removed from the industry chain and be unable to resume production for a long time.

Since the occurrence of industry chain risks is unavoidable and even unpredictable, we focus on the resumption of normal production after risks occur. In other words, the nature of industry chain security is not to completely avoid the occurrence of risks, but to effectively cope with the impact of risks. This implies that if we focus on industry chain risks caused by international competition, we need to examine China's irreplaceability in various industry chains. However, questions lie in how to measure a country's irreplaceability, to identify China’s competitive and uncompetitive industries, and those facing different industry chain risks.

1.1.2.1 Decentralization: Low-Tech Industries Face Horizontal Risks

Since joining the WTO, China has grown into the world's largest exporter over the past two decades. However, a breakdown of China's exports into primary products, low-tech manufacturing industries, and medium- and high-tech manufacturing industries,⁶ as depicted by the revealed comparative advantage (RCA) index⁷ (a relative indicator) and the export share⁸ (an absolute indicator) shows China’s current advantages mainly lie in low-tech manufacturing industries. In 2019, China’s RCA indexes for primary products, low-tech manufacturing industries, and medium- and high-tech manufacturing industries were 0.22, 1.27, and 1.11; and China’s shares of global exports for the three categories of industries were 3%, 17.4%, and 15.3%.

We take a further look at the RCA indexes of 97 industries classified by two-digit HS codes. Among the 18 primary product industries, China's RCA index was above 1 for eight industries in 2000, but for only three industries in 2019. Among the 58 low-tech manufacturing industries, China's RCA index was above 2.5 for 16 industries and above 1 for 34 industries in 2019, pointing to China’s strong comparative advantages in low-tech manufacturing industries. Among the 21 medium- and high-tech manufacturing industries, China is not very competitive, with its RCA index above 1 for only eight industries in 2019. China has relatively clear comparative advantages in the musical instruments, electrical machinery, appliances and equipment, and electric railway vehicle industries. However, China does not show comparative advantages in the aircraft, pharmaceutical, cosmetics, and automobile industries, and this situation has not improved much in the past two decades.

The export shares of the 97 industries also show China's current advantages mainly lie in low-tech manufacturing industries. Among the 18 primary product industries, China’s feather and down products industry had the highest export share at 22% in 2019. Among the 58 low-tech manufacturing industries, China's export share exceeded 20% for 28 industries, and was more than 50% for six industries, and reached about 80% for feather and down products and umbrellas in 2019. Among the 21 medium- and high-tech manufacturing industries, China's export shares in 2019 were not high; the industry with the highest export share was musical instruments (and related parts) at 25%.

In general, China's comparative advantages in export trade are mainly in low-tech manufacturing industries. However, the current comparative advantage does not mean that these industries are risk-free. In fact, we think these industries face horizontal risks caused by deglobalization. As mentioned earlier, China is a central link in global trade of these low-tech manufacturing industries. Therefore, the horizontal risk under trade frictions is decentralization—that is, reduced dependence on China as a central trade node. An important manifestation of horizontal risk is external demand decline and production capacity transfer.

1.1.2.2 China Faces Vertical Risks in Some High-Tech Industries and Primary Products

Although analysis from an export perspective confirms China's central position in global trade of finished products, China's advantages are smaller in trade of intermediate products. In our opinion, the more complex the division of labor for an intermediate product, the smaller China's advantage is in trade. Analysis based on the global value chain position index shows China as a whole is at the end of the global industry chain. The higher the index, the further the country is from end-consumers. Russia and Australia have high indexes because they export raw materials such as minerals and oil & gas resources. Japan and the US have higher indexes because they are in the midstream and upstream links such as R&D. China is in the downstream of the industry chain, and its imports are constrained by countries that export basic resources, key technologies, and key equipment (Fig. 1.4). It is necessary to further analyze China's industry chain risks in light of the degree of dependence on imports.

Fig. 1.4

2019 global value chain position index. Note The higher the global value chain position index, the closer the country to the upstream of the global value chain. The size of the bubble represents the size of the domestic value added absorbed by foreign countries in the country's total exports. Source Gross Trade Accounting: Official Trade Statistics and Measurement of Global Value Chains (Wang Zhi et al. 2015), ADB MRIO database, CICC Global Institute

China accounted for 11.2% of global imports in 2019 and was the world's second-largest importer, only behind the US (14%). China’s median share of global imports in the 1,242 industries classified by four-digit HS codes in 2019 was 4%, an increase of 2.2ppt from the level in 2000. Among imports in primary products, low-tech manufacturing industries, and medium- and high-tech manufacturing industries, China had the highest import share for primary products, while the US had the highest import shares for both low-tech manufacturing industries and medium- and high-tech manufacturing industries (Fig. 1.5).

Fig. 1.5

Import shares of China and G7 countries, 2019. Note Import share = the country's import value of the category of goods / global import value of the category of goods. Source UN Comtrade, CICC Global Institute

We note that China's overall import share in 2019 was higher than the 75th percentile of its import shares among 1,242 industries (Fig. 1.6). This means that China's import structure is polarized—more than 75% of industries have an import share below average, while the heavy import dependence of a small number of industries significantly increases the overall import dependence.

Fig. 1.6

Import share distributions of China and G7 countries by industry. Note Import share = the country's import value of the category of goods / global import value of the category of goods. Source UN Comtrade, CICC Global Institute

Our definition of heavy import dependence includes three dimensions—namely import value, share of global imports, and concentration ratio of top-four sources of imports (CR4) (Fig. 1.7). The larger the first two indicators, the higher the dependence of the industry chain is on imports; the higher the third indicator, the higher the dependence on imports from one or a few markets and the more susceptible to international competition or natural disasters. Considering the size of China's economy, we exclude industries with an import value of less than US$2bn from the 1,223 industries classified by four-digit HS codes, and screen industries with high import dependence.

Fig. 1.7

Screening framework for industries with high import dependence. Note Screening is based on the 1,223 industries classified by four-digit HS codes. China accounted for 16% of global GDP in 2019, and we consider an import share above 32% (double the GDP share) as high. We refer to the market concentration standards, and an import source CR4 above 75% means that the market is highly oligopolistic. CR4—concentration ratio of top-four sources of imports. Source CICC Global Institute

Among China’s 207 primary product industries, 24 industries (11.6% of the total) are identified as having high import dependence, especially mineral resources such as petroleum crude oil, iron ore, petroleum gas, copper ore, and oil crops such as soybeans (Fig. 1.8). Some products (such as iron ore, soybeans, and palm oil) have an import source CR4 above 90%.

Fig. 1.8

Among primary products, mineral resources and oil crops have heavy import dependence. Note The size of the circle represents the value of imports of various products by the Chinese mainland in 2019; import share = the country's import value of the category of goods/global import value of the category of goods, both in 2019. Source UN Comtrade, CICC Global Institute

Among the 609 low-tech manufacturing industries, only 35 industries (5.7% of the total) are identified as having high import dependence, mainly rough processing of primary products such as refined copper and copper alloys (Fig. 1.9). These industries are not technologically complex, and their import concentration is relatively low. China is unlikely to be technically constrained by other countries in low-tech manufacturing industries. The more important consideration is ensuring the import security of upstream mineral resources.

Fig. 1.9

Among low-tech manufacturing industries, primary metal smelting products have heavy import dependence. Note The size of the circle represents the value of imports of various products into the Chinese mainland in 2019; import share = the country's import value of the category of goods / global import value of the category of goods, both in 2019. Source UN Comtrade, CICC Global Institutem

To sum up, China’s disadvantages lie in primary products such as petroleum, soybeans, and iron ore, as well as medium- and high-tech manufacturing industries such as integrated circuits, integrated circuit manufacturing equipment, motor vehicles, and pharmaceuticals. If these industries with high import dependence experience a cutoff in overseas supply, their domestic production could be disrupted. Since such risk comes from the midstream and upstream, it is called “vertical risk.”

1.1.3 Technological Innovation: Key to Improving China’s Industry Chain Security

In our opinion, the fundamental reason for the low irreplaceability of China lies in insufficient R&D investment. This issue exists not only in high-tech industries, but also in low-tech manufacturing industries where China has advantages. This is a common problem in China's manufacturing sector. The capital intensity is higher than the R&D intensity in almost all of China’s industries, and manufacturing companies compete more by expanding production capacity than by increasing R&D (Fig. 1.10). This may cause a waste of resources and excessive competition in low-end industries as well as insufficient R&D investment and weakened incentives for companies to move up the value chain. Therefore, increasing R&D intensity and accelerating technological innovation are not only needed in high-tech industries to enhance value chain security, but in various other industries as well.

Fig. 1.10

China's manufacturing sector has large capital expenditures but insufficient R&D investment. Note Data of listed manufacturing companies in the countries in 2020. R&D (capital) intensity = R&D (capital) investment/sales revenue. Source Factset, CICC Research Manufacturing Team

Regarding primary products, China relies heavily on imports of important raw materials (such as oil) that pose vertical risks. In general, as primary products are natural resources, the solution to vertical risk includes economic means such as diversifying import sources, building domestic stock, and non-economic means. However, these are only short-term measures with questionable effectiveness. We believe China needs to rely on technological progress in green energy and other areas to solve the problem of heavy import dependence on primary products such as oil.

As mentioned earlier, China is facing horizontal risk caused by deglobalization in low-tech manufacturing industries where it has comparative advantages. We emphasize that although some companies have relocated their production lines to Vietnam, the Philippines, and other countries to avoid tariff issues since 2018, more global companies have adhered to a China-based strategy. This shows trade frictions alone are not enough to change China's comparative advantages. The key horizontal risk lies in technological progress. A study shows that every robot added in the US manufacturing industry replaces 3.3 workers.⁹ The widespread use of industrial robots appears to be slowly bringing manufacturing back to the US, and the size of manufacturing investment in the US has increased steadily in the past decade. Against this background, developing countries that rely on demographic dividend to serve as the “world's factories” may face the possibility of export production capacity being substituted. In our opinion, for China to fundamentally eliminate horizontal risks faced by its low-tech industries, it is vital to enhance irreplaceability through technological progress.

1.2 Assessment of Innovation Activity in China: From the Perspective of Innovation System

As previously discussed, we believe the key to strengthening irreplaceability is to improve R&D investment. However, R&D cannot function alone, and innovation activities also need competent researchers and a supportive policy environment. Moving from a global perspective to innovation in China, we seek to provide a full picture of China’s innovation status quo based on the structure of its innovation system. The innovation system of a country usually consists of three parts: Technological innovation input, technological innovation environment, and technological innovation output¹⁰ (Fig. 1.11). This section focuses on evaluating the inputs and outputs of technological innovation in China. Innovation inputs include capital investment and human resource investment, of which capital investment features R&D. As discussed in section 1, R&D is at the core of accelerating innovation. It comprises creative and systematic work undertaken in order to increase the stock of knowledge and to devise new applications of available knowledge. Talent is also essential to innovation, and China benefits from its large talent base. Through knowledge and production innovation activities, capital and human resource inputs convert into innovation outputs, including academic publications and patents, among others.

Fig. 1.11

Key indicators in a national science and technology (S&T) innovation system. Note Technological innovation mechanism and environment in the figure based on WIPO Global Innovation Index framework. Source OECD, CICC Global Institute

1.2.1 Technological Innovation Inputs: R&D Expenditure and Human Resources

1.2.1.1 Is China’s R&D Investment Sufficient?

China’s R&D investment has increased rapidly since 2000, and currently is the second largest in the world. The country’s R&D intensity has also increased year by year and has reached 2.4%, which is close to the average of advanced economies and much higher than the average of developing countries. However, is China's R&D investment large enough?

An important factor that affects R&D intensity is a country’s economic performance, which can be measured by per-capita GDP. Data from 43 countries in 2017 indicates that R&D intensity is positively correlated with per-capita GDP (Fig. 1.12). Given China's current per-capita GDP, the country’s R&D intensity seems high and is close to the level of many developed economies. As major technology powerhouses' R&D intensity reached 2.23% in 2019, the same level as China's current R&D intensity, their per-capita GDP at least exceeded US$10,000 or even US$20,000. However, China's per-capita GDP is only US$8,342.3 in 2020 (based on constant US dollar in 2010).¹¹

Fig. 1.12

Relationship between R&D intensity and per-capita GDP in sample countries worldwide, 2017. Note Per-capita GDP is calculated based on US dollar value in 2017. Source OECD，World Bank, CICC Global Institute

We believe that it is imperative for China to increase its R&D intensity in order to achieve long-term economic development targets. In addition, raising R&D intensity has been shown to be an important way for less developed countries to catch up with advanced economies. For example, R&D intensity in Japan, South Korea, and northern European countries exceeded that of the US amid their rise, marking a sharp contrast with the low R&D intensity in Latin American countries and other countries in Asia and Europe.¹² We note that China, as a large country that is attempting to catch up with developed economies, is likely to become the first country whose R&D intensity exceeds 3% when its per-capita GDP is less than US$30,000.

However, China faces structural problems in R&D spending. Companies have accounted for an increasingly large proportion of China's R&D sources since 1990. They replaced the government as the largest source of R&D funding in 1997, and their share is now at 76%. In contrast, the government's share of R&D investment trended downward to around 20% in 2013 from the late 1990s. Although companies contribute a large part of China's R&D spending, their R&D intensity remains low and they are reluctant to invest in basic research.

Corporate R&D intensity refers to the ratio of companies' R&D spending to their revenue. We find that a country's R&D intensity exhibits a clear linear correlation with corporate R&D intensity.¹³ China's corporate R&D intensity was only slightly above 2% in 2017. We believe this is an important reason why China's overall R&D intensity is still lower than technology powerhouses. The industry of instrument and apparatus is the only one in China that has a corporate R&D intensity over 3%. The R&D intensity is only 2.19% for the computer, communication, and electronics industries, and even lower than China's overall R&D intensity.¹⁴ In contrast, R&D intensity in multiple industries in the UK has exceeded 3%, and it even surpassed 5% in a number of industries in the US.¹⁵

China’s spending on basic research is insufficient, in our opinion. In China, R&D as a percentage of GDP is comparable with that in the US, but the vast majority of R&D investment goes to Development (D)—i.e., the development and promotion of products and technologies—while Research (R) accounts for a much lower proportion in total R&D investment than in the US.

Spending on experimental development as a percentage of China’s total R&D spending reached 82.7% in 2018, exceeding that of US, Japan, South Korea, and Israel. However, the share of China’s spending on basic research is much lower. In contrast, the share of the spending on basic research is above 10% in the UK, US, South Korea, and Japan, reflecting their strong emphasis on basic research. During 2016–2020, China's spending on basic research grew at an annual average rate of 17%, and the share of basic research spending in Chinese R&D investment has trended upward.¹⁶ Nevertheless, the share was only 6.03% in 2018. The insufficient spending on basic research results in a weak foundation for technological development and poor capabilities in radical innovations and inadequate foundations for technological development.

One critical reason for China's insufficient spending on basic research is the lack of corporates’ significant investment in this area. In 2018, corporate R&D spending accounted for 76% of Chinese R&D spending, but basic research made up only 0.22% of corporate R&D spending. Consequently, the share of companies spending on basic research was only 3% in China, far below the average share of 28% in the US.

Chinese companies are reluctant to participate in basic research.¹⁷ During the years when China was a planned economy, companies mainly focused on production, while universities and research institutes were in charge of technology research.¹⁸ Consequently, it was long believed that basic research should be done by universities and research institutes instead of companies. Chinese firms are unwilling to participate in basic research as they believe they can acquire technologies through technological transfer.

Meanwhile, Chinese companies historically have not been actively involved in national science and technology programs. In particular, National Natural Science Foundation's projects and the “973 program” that both focus on fundamental science have been rarely undertaken by companies. China carried out its national key lab program as early as in 1984, while it did not create its corporate key lab scheme until 2006. Although China now emphasizes the role of companies in key national R&D projects, companies remain primarily focused on applied research and universities or research institutes are still in charge of basic research tasks.

In addition, Chinese demand for basic research is weak due to its position in the global industrial division. As the world's largest manufacturer of finished goods, China focuses on processing and assembly procedures in the global value chain. Low-tech manufacturing industries make up the largest proportion in its exports.¹⁹ Consequently, domestic companies have tepid demand for R&D, especially for basic research. According to a survey by McKinsey in early 2021, R&D activities targeting product cost reduction made up two-thirds of R&D spending by Chinese companies.²⁰

Private companies have lower R&D intensity than state-owned enterprises (SOE). In particular, their spending on basic research is relatively low. In 2019, private companies accounted for 39.6% of total R&D spending from Chinese companies and 71% of domestic R&D spending on high-tech areas.²¹ However, their R&D intensity and spending on basic research are both relatively low. In high-tech areas, SOEs had R&D intensity of 4.9%, while the R&D intensity of all Chinese companies was only 2.7%. We believe this suggests that SOEs focus more on basic research and that their R&D output is stronger than other firms.²²

1.2.1.2 Large Talent Base, but Larger Room for Improvement

China has the largest talent base in the world. According to data from the Organization for Economic Cooperation and Development (OECD), China had a total of about 1.87mn researchers in 2018, more than 1.43mn in the US, 680,000 in Japan, 430,000 in Germany, and 340,000 in India.²³ Regarding higher education and professional knowledge, China leads the world in the number of science, technology, engineering, and mathematics (STEM) graduates with a bachelor's degree or above—China’s 1.79mn graduates in 2018 represented three times that of the US, nine times that of Germany and the UK, 12 times that of Japan, and 13 times that of South Korea.

However, China still lags in terms of talent per capita. Data from the OECD and the World Bank showed that the number of researchers per 1,000 workers in China in 2018 was only one-sixth that of South Korea, one-fifth that of Singapore, and one-fourth that of the US, Germany, and Japan, and slightly lower than the level commensurate to GDP per capita. Compared with the US, major developed countries in Europe, Japan, and South Korea, China had a lower number of years of education completed per capita in 2018, which was also below the average level corresponding to its GDP per capita (Fig. 1.13).

Fig. 1.13

Number of years of education completed. Note 2018 data. Source OECD, World Bank, CICC Research

Lack of high-quality talent is particularly prominent in China's key industries. According to data from Macropolo, 59% of the top artificial intelligence (AI) workers choose to work in the US, followed by China (11%), Europe (10%), Canada (6%), and the UK (4%). Tsinghua University and Peking University are the only two Chinese institutions that are included among the top 25 AI research institutions in the world, ranking No. 9 and No. 18, respectively.²⁴ In the integrated circuit (IC) industry in China, holders of masters and doctoral degrees account for only 18% and 1% of total employees. The White Paper on Talent in China’s Integrated Circuit Industry (2019–2020) published by the Shenzhen Semiconductor Industry Association, predicted demand for talent in the industry would reach 720,000 around 2021. However, there were only 199,000 graduates of IC-related majors from domestic colleges and universities in 2018.

From our perspective, talent resources are unevenly distributed among industries in China. Judging from A-share listed companies, the three industries with the highest proportion of employees with a bachelor’s degree or above are banking (84.1%), non-banking finance (82.4%), and diversified finance (65.8%). These three industries do not have high demand for R&D but have more graduates with higher education, which may represent a misallocation of resources, in our view. Many talented workers are more inclined to be engaged in the financial industry rather than in R&D. This is related to the fact that the finance industry has a higher salary level than other industries.

We believe three supply-side challenges constrain the advancement of high-quality basic education, resulting in inadequate innovative personnel: The uneven distribution of high-quality basic education in China directly reduces the quantity of human capital. The strong incentive system of exam-oriented education has resulted in excessive investment in exam-oriented skills. Research-oriented universities should play an important role in turning human capital into innovative personnel, but their development potential is constrained by educational administration factors.

First, the quality of basic education in developed regions of China is high, but the distribution of spending on basic education is uneven among regions. Education expenditure as a percentage of GDP is 4.6% in China, and basic education expenditure accounts for 70% of total education expenditures. Both indicators are mediocre compared with those in major industrialized countries. However, the education expenditure per ordinary primary school student in developed regions is generally higher than that in underdeveloped regions. Data from the China Educational Finance Statistical Yearbook shows that the highest education expenditures per ordinary primary school student were in Beijing (at Rmb39,000), Shenzhen (at Rmb37,000), and Shanghai (at Rmb34,000) in 2019, while Henan had the lowest expenditure at only Rmb7,953. The coefficient of variation in public education expenditure per ordinary primary school student in China’s provinces is 0.4, higher than that in public education expenditure per K-12 student in the US (0.3).

The uneven distribution of basic education spending also exists within regions. In 2018, the education expenditure per student at local rural primary schools was Rmb11,827, 7% lower than ordinary primary schools at Rmb12,737. The linkage between housing and education could exacerbate inequality in education opportunities, in our opinion. In addition to basic residential functions, housing also corresponds with basic public services. Data from ke.com shows that in major cities with high housing prices, housing near key schools has large premiums. High housing prices have become a threshold for high-quality education resources, exacerbating the uneven distribution of education resources.

Second, strong incentives for exam-oriented education has led to excessive investment in exam-oriented skills. The exam-oriented selection mechanism is a strong incentive system that increases the motivation of students, parents, and teachers and enhances cognitive abilities of students. This is a crucial reason for Chinese students ranking among the best in the Programme for International Student Assessment (PISA) test. However, such a one-dimensional strong incentive system could easily lead to a “prisoner's dilemma” situation where students place too much effort into improving exam-oriented skills instead of increasing human capital, resulting in a waste of resources.

Third, research-oriented universities should be given more space for innovation, in our opinion. Research-oriented universities turn human capital into innovation, and high-quality research-oriented universities can greatly increase the proportion of innovators. Bloom et al. (2021) found that disruptive technology centers are more likely to appear in areas with universities and highly skilled labor.²⁵ China has a large number of universities but a relatively small number of high-quality research-oriented universities. In our view, one reason is that the development of research-oriented universities is influenced by administrative factors.

China's spending on higher education is relatively low, and government appropriation plays a significant role. Based on data from the China Educational Finance Statistical Yearbook, China's education spending accounted for about 4% of GDP in 2018, lower than the proportion in the US and the UK (about 6% each), according to the OECD. In addition, government appropriation accounted for 67% of China's spending on higher education, much higher than the proportion in the US (35%) and the UK (26%).

Government appropriation to higher education ensures equity in education, but it also imposes constraints on higher education institutions. Compared with the US, higher education in China is cheaper and more inclusive. However, colleges and universities in China also face more constraints when using government funding. For example, 82.5% of scientific and technological workers think researchers with certain administrative titles are more likely to obtain research funding.²⁶

A well-structured talent system both expands the supply of talent from the domestic education system and attracts cutting-edge talent through its favorable environment around the world. In the US, immigrants account for 18% of all talent. Further, 58% of all global immigrant talent chooses the US (Fig. 1.14), and 42% of US PhD graduates in STEM majors in 2019 were students from other countries. Such immigrants are often the top scientific and technological workers trained by their home countries.

Fig. 1.14

Number of immigrant talent and emigrants of select countries, 2001–2010. Source Fink, C, E Miguélez, and J Raffo (2013), “The global race for inventors”, WIPO Economic Research Working Paper, CICC Research

In comparison, China sees a large outflow of talent, especially in STEM majors. China has a large number of students abroad. In 2017, there were about 928,000 Chinese students abroad, accounting for about 17.5% of global students abroad. The proportion of Chinese students studying STEM majors in the US has increased, while the proportion of STEM graduates returning to China is relatively low. In the 2019–2020 academic year, there were about 181,000 Chinese students majoring in STEM in the US, accounting for nearly 50% of the total number of Chinese students studying in the US. This proportion has been increasing since 2012. According to the Annual Report on the Development of Chinese Students Studying Abroad (2020), students majoring in economics and management accounted for about 46% of returning graduates, while students majoring in science and engineering accounted for 31%.

China's attractiveness to top talent needs to be improved. Among students earning doctoral degrees in the US, Chinese students take up a much larger proportion than students from other countries. Data from the US National Science Foundation showed about 6,300 Chinese students received doctoral degrees in the US in 2019, three times the number of Indian students, 13 times the number of South Korean students, and 26 times the number of Japanese students. Also, 80% of Chinese students who have earned a doctoral degree in the US are willing to stay there, much higher than the proportion of 51% for Japanese students and 65% for South Korean students. Data from Macropolo shows only 31% of the top AI researchers working in the US are from the US, and the remaining 69% are international students who stay in the US after graduation. China accounts for the highest proportion of such students (27%), with all of Europe accounting for 11%. In 2020, 88% of Chinese PhD students in the US focusing on AI chose to work in the US after graduation, while 10% chose to return to China.²⁷

Excessive competition may increase the difficulty of attracting talent. The talent market is first of all a labor market, with relatively fixed supply. For such a market, using subsidies and other means to stimulate demand may lead to an increase in wages rather than an increase in talent.²⁸ Market entities may make excessive commitments in order to attract talent amid competition. Such a phenomenon would intensify the information asymmetry in the labor market, weaken the role of the signaling mechanism in the labor market, reduce the efficiency of matching between talent and employers, and even lead to adverse selection in the talent market.²⁹ This may reduce the number of workers a college or university attracts in the end due to market friction. For example, some colleges and universities may over-promise research funding, salary, benefits, and pace of career advancement when attracting talent, and fail to deliver on their promises. This weakens the efforts to attract talent and increases the difficulty of attracting talent in the future.

1.2.2 Technological Innovation Output: Five Patterns of China’s Progress in Playing Catch-Up in Knowledge Innovation Over Three Decades

Technological innovation output converted from R&D and talent inputs could most directly reflect a country's scientific research strength in the past. It can be regarded as the “transcript” of the country's achievements in technological innovation. Moreover, achievements in technological innovation could also guide a country's technological innovation input and construction of an innovation system. By comparing the achievements of technological innovation activity in various countries, we could gain insights and draw lessons from technological innovation activity in the past, which gives direction to the improvement in innovation system and resource allocation. Therefore, a review of previous achievements in technological innovation in a results-oriented approach provides guidance for improving China's technological innovation activity.

Human society’s technological innovation activity can be divided into two types: Knowledge innovation and production innovation. In a nutshell, knowledge innovation refers to human activity involved in developing basic science through scientific research and applying new scientific knowledge. By contrast, production innovation refers to the combination of knowledge production and production of goods and services with the help of innovative thinking. Compared with production innovation, knowledge innovation not only is a prerequisite for humanity to understand and change the world, but also determines the breadth and depth of production innovation. In addition, while production innovation tends to directly affect production in the near term, knowledge innovation exerts a greater impact on production activity in the future. Therefore, in this section, we place our focus on knowledge innovation in China’s technological innovation activity. Production innovation will be examined in detail in the following chapters.

Among the achievements that can be studied, academic publications and invention patent applications are recognized by international institutions³⁰ as key indicators of achievements in knowledge innovation in various countries due to their good comparability, objectivity, and high relevance to science and technology. In order to present China’s achievements in knowledge innovation since the 1990s in a comprehensive way, we refer to the Microsoft Academic Database (MAD), Nature Index, SCImago Journal & Country Rank, and other databases that cover about 300mn academic papers published in nearly 50,000 journals in 27 disciplines, e.g., mathematics, physics, chemistry, life sciences, environmental sciences, materials science, and medicine. We also use databases from the World Intellectual Property Organization (WIPO) and relevant national patent offices to analyze more than 3mn PCT patent applications³¹ in electrical engineering, mechanical engineering, chemistry, instruments, and other fields, which can be further divided into 35 technology fields under the above five sectors.

We summarize the evolution of China’s technological innovation output in the past three decades and highlight five patterns of China’s knowledge innovation. We hope that the five patterns can help us better understand the pros and cons of China’s scientific and technological innovation activity in the past, and provide clues for improving China’s technological innovation ecosystem in the future.

1.2.2.1 Pattern 1: Catching Up in Quantity, and Narrowing Gap with Developed World in Quality

China was in the past a latecomer but is now taking the lead in academic publications, especially in STEM. The number of academic articles published from China each year accounted for less than 3% of the global total by 1995, far behind the 30% of the US. However, after nearly 30 years of catch-up, China saw its annual academic article publication output surpass that of the US to become the world’s largest contributor of academic articles, accounting for about 16% of the world’s total in 2020. Looking ahead, we believe that China may further enhance its advantage against the US in terms of annual academic article publication output.

Among all subjects, materials science, chemistry, and medicine were ranked the top three in China regarding the number of publications. Compared with the beginning of the 21st century, China’s research article publications were increasingly concentrated on STEM fields over 2015–2020. The share of publication output of almost all STEM subjects in total publications over 2015–2020 was higher than that over 2000–2005.

The number of international patent applications that China filed through the Patent Cooperation Treaty (i.e., PCT patents)³² in the 1990s was also very small. However, from 2000 to 2019, the average annual growth rate of the number of PCT patent applications filed by China was close to 26%,³³ versus 2%³⁴ in the US during the same period. This drove China to surpass the US to become the top filer of PCT patent applications in 2019. With the growing number of PCT patent applications filed by China, China’s share of the world’s total number of PCT patent applications has also trended upward year by year. From 2000 to 2020, China's share of the world’s annual PCT patent applications expanded to 25% from less than 1%, while that of the US dropped to 21% from 41%. Specifically, the increase in the number of China’s patent applications has mainly been driven by the electrical engineering sector. So far, patents in the electrical engineering account for about half of China’s total PCT patent applications.

Some people believe that the rapid increase in the number of China's academic publications and PCT patent applications is driven by quantity-oriented policy incentives,³⁵ and that such policies may negatively affect the quality of China’s knowledge innovation. However, multiple indicators show that China has also made significant progress in the past 20 years in terms of the overall quality of knowledge innovation, and that the “quality gap” between China and developed countries has been narrowing.

A higher number of citations indicates a greater level of recognition a paper receives from peers. Therefore, the number of citations is often regarded as an important indicator of the quality of an academic article. Compared with the number of citations in different disciplines in the US during the same period, we have found that the average number of citations for articles from China increased notably in the past 20 years. Among the articles published over 2000–2002 (by time of publication), the average number of citations for most China’s articles was significantly lower than that of US ones.³⁶ Compared to the citation count ratio over 2000–2002, amid the increase in the total number of publications, the average number of citations for articles published in China compared to articles published in the US in various disciplines also increased over 2017–2019. That said, based on rankings of all subjects in terms of citation rate, although China still lags behind the US in STEM fields (Fig. 1.15), China has made progress in research quality over the past 20 years. In addition, from the perspective of patent, we have also found an increase in the number of China’s PCT patent citations, which shows that the quality of knowledge innovation in China has been improving, and that the gap between China and the US has been narrowing.

Fig. 1.15

The average number of citations of research articles in most disciplines in China increased significantly versus US counterparts. Source SCImago Journal & Country Rank, CICC Global Institute

1.2.2.2 Pattern 2: China Still Accounts for a Small Share of Global Core Knowledge Network; Originality Remains a Weakness

Although the output of human knowledge can be random to a certain extent, it is rarely produced in isolation. The information on citing and cited publications lays a foundation for the establishment of a global knowledge network. Through such a knowledge network, we can gain an overall view of the “inheritance system” of human knowledge, and understand the importance of each innovation activity. This is of great significance for understanding the time, place, and technical field of innovation activity, as well as for exploring the evolutionary patterns behind such activity.

We extract information on research articles published and cited since 1995 from Microsoft Academic Graph as well as all PCT patent information since 1978 from EPO Patstat, and present their citation networks. In these citation graphs, each point (node) represents a paper or a patent, and the distance between points depends on the citation relationship between them. The closer the two points are in the network, the tighter their citation relationship is.³⁷ Areas with higher density in the network represent core fields of research, while areas with low density correspond to general knowledge or technologies applied in various disciplines or industries.

Figure 1.16 shows the citation networks of academic literature published since 1995. Some highly interrelated papers such as papers in computer science, biology, etc., have developed into relatively independent areas. In particular, in the computer science citation network (in yellow), most of the literature keeps expanding within its own knowledge system, and some mathematics and social sciences literature are derived on the right side of the computer science field. Engineering (in red) demonstrates a different scenario. Although it ranks No. 2 among all fields in terms of number of papers, its citation network intersects with a number of other disciplines, including computer science, materials science, and mathematics. This shows that backed by a large amount of academic literature in other fields, engineering continues to achieve innovative research output while offering support to innovative research in other disciplines. There are also some disciplines that are closely related to each other. They form an interwoven citation network around the basic disciplines. For example, materials science is relatively closely related to chemistry, medicine, and biology.

Fig. 1.16

Paper citation networks (1995–2020). Note The figure shows the mutual citation relationships of nearly 1.1 mn academic documents. The data includes research articles published from 1995 onward with complete country and discipline information. Source Microsoft Academic Graph, CICC Global Institute

As shown in Fig. 1.17, from the perspective of discipline structure, the citation networks of papers from China have mainly been distributed in the fields of computer science, materials science, and biology in recent years, while in other fields, there are only sporadic connections. In the citation network, papers closer to the core position of the citation cloud of a field are more frequently cited. However, most of the citation networks related to China in the figure are located at the edge of the corresponding discipline areas, and the density is low. This to a certain extent indicates that the proportion of papers from China in the core positions remains relatively low. In addition, in the citation network, we also note that currently, China's academic literature tends to concentrate in applied science disciplines, while literature in basic science disciplines is insufficient.

Fig. 1.17

Distribution of China’s academic papers (1995–2020). Note 1) This figure shows the mutual citation relationships of nearly 1.1 mn academic documents. The data includes research articles published from 1995 onward with complete country and discipline information. 2) Red dots represent papers from China and gray dots represent those of other countries. Source Microsoft Academic Graph, CICC Global Institute

Figure 1.18 shows the global citation network of PCT patents since 1978. Patents in the electrical engineering and chemistry sectors are relatively concentrated in terms of distribution, and they have little overlap with other sectors, indicating that there are more citations between these two sectors. In comparison, the mechanical engineering and instruments sectors are at the core of the citation network. They are intertwined with each other and have more connections with other sectors. This suggests that the mechanical engineering and instruments sectors play a role in providing extensive general technology and product support for other sectors.

Fig. 1.18

PCT patent citation network for the technology sector (1978–2020). Note The above figure shows the mutual citation relationships of nearly 2.97 mn PCT patent families. The data includes all PCT patent families filed from 1978 onward, and these patent families were cited by another PCT patent family at least once. Source EPO Patstat, CICC Global Institute

Figure 1.19 depicts the positions of China’s PCT patent applications in the global patent citation network. We can clearly see that China's PCT patents are mainly concentrated in the peripheral area of the electrical engineering field, with darker colors only in some areas. The distribution of China's patents is more sporadic in other fields. This is largely in line with our previous discussion on academic publications.

Fig. 1.19

Distribution of China’s PCT patent applications (1978–2020). Note 1) The above figure shows the mutual citation relationships of nearly 2.97 mn PCT patent families. The data includes all PCT patent families filed from 1978 onward, and these patent families were cited by another PCT patent family at least once; 2) The red dots in the above figure represent China’s patents, and the gray dots represent those of other countries. Source EPO Patstat, CICC Global Institute

Originality has self-evident importance for science, but objectively measuring originality poses a formidable challenge. We introduce a new indicator to measure the originality of a work in knowledge innovation.³⁸ Specifically, if the citation relationship between the subsequent work citing an achievement in knowledge innovation and its reference is not close, the originality index of the achievement is relatively high. Compared with the traditional method of using citation count to measure its quality, the originality index also takes into account the previous citation relationships, and could thus reflect the uniqueness, innovativeness, and importance of the achievements in a more objective and comprehensive way.

We use data from Microsoft Academic to calculate the originality index of China's and US academic papers from 2010 to 2015. We have found that the level of originality of papers from China is similar to that of US papers in materials science and physics, but lags behind the US in most other disciplines, especially in medicine, computer science, and engineering (Fig. 1.20).

Fig. 1.20

China lags the US in terms of originality in most STEM disciplines. Source Microsoft Academic, CICC Global Institute

In terms of the originality indexes of PCT patents (2015–2017 filing years), the US outperformed China in almost all fields. Among the five main technology sectors, the originality of China’s patents in the electrical engineering discipline was relatively high. It is noteworthy that the level of originality of China's patents was slightly higher than that of their US counterparts in the semiconductor field. One possible reason is that due to limited technological and research collaboration, China had to carry out independent R&D in the semiconductor field, which to a certain extent increased the originality of China's patents in this field.

1.2.2.3 Pattern 3: “Strong Industry, Weak Academia” in Conversion of Knowledge Innovation, Scarcity of Frontier Tech Companies in the Market

The conversion of achievements in knowledge innovation can take two forms: The conversion of basic research innovations into commercial patents, which could be measured via the conversion efficiency of basic research results³⁹; and the adoption of patented technology in the industry, which could be measured via the rate of patent commercialization.⁴⁰

In China, basic research work is mainly undertaken by universities and research institutes. From the perspective of conversion of basic research results into commercial patents, China is still a laggard compared to the US and Japan. In the internet information sector and the manufacturing and engineering sector, basic research provides little support for related commercial patents in China (Fig. 1.21), indicating that the connection between basic research and the development of commercial technologies is weak.

Fig. 1.21

China lags behind the US and Japan in conversion efficiency of basic research (2020). Source Evaluation of transformation efficiency of basic research results in China's key technical fields,Footnote

Wu, F., Li, Y., Miao, H., Huang, L. (2021). Evaluation of the transformation efficiency of basic research results in China's key technical fields, Scientific Research.

CICC Global Institute

The overall commercialization rate of China’s patents is not low. However, it lags well behind developed countries in the commercialization of commercial patents from universities and research institutes. Similar to enterprises in many developed countries, Chinese enterprises also play a dominant position in patent application. The number of PCT patent applications submitted by Chinese enterprises accounted for more than 80% of China’s total patent applications in 2018. The patent commercialization rate of Chinese enterprises is about 45%, close to the rates in developed countries, and it raises the overall commercialization level of Chinese patents. However, the commercialization rate of patents from Chinese universities and research institutes was 3.8% in 2018, significantly lower than that in the US (50.4%),⁴² which could be a drag on the average commercialization rate.

Scarcity of frontier tech companies might be an impediment to application of knowledge innovation achievements. China accounts for relatively high proportions of global academic publications and patent applications for all 11 frontier technologies⁴³ defined by the United Nations Conference Trade and Development (UNCTAD). For example, as of 2018, China’s academic papers in the 11 frontier fields such as artificial intelligence (AI), Internet of Things (IoT), big data, and blockchain accounted for 13% of the global total, versus the 21% in the US; China’s patents accounted for 22% of the world’s total, versus the 30% in the US, indicating that China performed relatively well overall.

However, world-leading enterprises in frontier technology areas are scarce in China, with drones and solar photovoltaics (PV) being the only exceptions. This may have something to do with the fact that knowledge innovation in frontier fields has not yet helped relevant enterprises enhance their competitiveness in China. Another explanation is that relevant Chinese enterprises may not yet possess enough competency in cutting-edge fields, which hinders the conversion of achievements in knowledge innovation. In either case, it shows that China faces constraints in the application of knowledge innovation in frontier fields.

1.2.2.4 Pattern 4: Behind High-Quality Innovation is Comparative Advantage, Which is Both a Driving Force and a Constraint

Measured by achievements in research, China stands out in some areas, while it slightly underperforms in others. In areas featuring a large quantity of high-quality research and high conversion or commercialization rates, market forces are usually at play. In particular, the areas in which China performs well in knowledge innovation activity are often areas in which China shows comparative advantages in international trade.

As we mentioned earlier in this chapter, China’s knowledge innovation in the electrical engineering industry shines in terms of both quality and quantity. Meanwhile, international trade activity in China's electrical engineering-related industries is also strong. To explain this correspondence, we have found that the “hidden comparative advantage” plays an important role. According to the OECD’s classification for manufacturing industries (2016),⁴⁴ technology products can be divided into four categories, namely high, medium-high, medium, and medium-low technology products. High-tech manufacturing mainly includes aircraft, spacecraft, and related machinery; computer, electronic, and optical products; as well as pharmaceuticals. China has not yet fundamentally changed the fact that its comparative advantage mainly lies in labor-intensive sectors. Therefore, high-tech manufacturing does not seem to be in line with the country’s comparative advantages. However, if we categorize the medium-high-technology industries defined by the OECD based on their capital intensity and technology intensity, we would find that computer and communications shows the most substantial comparative advantage among the high-tech industries in China.

With reference to the OECD’s definition, we draw a scatter plot of the six medium-high technological manufacturing industries in China and the US based on the intensity of capital and technology (Fig. 1.22). In contrast to the general perception of high-tech industries, in the US, the capital intensity of computers, communications, and other electronic equipment industries is lower than that of the aerospace and pharmaceutical industries. Moreover, the capital intensity of the computer and communications industry is even lower than that of some medium-high technology industries such as the automobile and chemical raw materials and products industries. The computer and communications industry is relatively labor-intensive among high-tech industries, which perfectly fits the comparative advantage of China in terms of labor endowment. Therefore, China has formed a virtuous circle between knowledge innovation and corporate profits in computer, communications, and electronic equipment.⁴⁵

Fig. 1.22

China shows comparative advantages in computer and communications fields. Note 1) Capital labor ratio refers to net value of fixed assets of the industry/number of employees (Rmb10,000/person, based on data released by Shanghai Academy of Social Sciences), and technology intensity refers to the proportion of researchers in the industry; 2) US indicators in gold use 2016 data. Chinese indicators in red use 2017 data. Source NBSC, US National Science Foundation, US Bureau of Economic Analysis (BEA), CICC Global Institute

In fact, the positive feedback relationship between comparative advantages and innovation capabilities in China's electrical engineering field is not an isolated case. Following the OECD’s classification standard of manufacturing industries (2016), we calculate the revealed comparative advantage (RCA) index in terms of patents and exports of China and the US under the four technology categories in 2000 and 2019, and have found the RCA of patents and the RCA of international trade in corresponding fields basically maintained a clear positive correlation. For example, the comparative advantage of the US in international trade lies in the high-tech field, which remains unchanged over the past two decades. Correspondingly, the comparative advantage of its patent applications is also notable in the high-tech field. The sectors in which China had comparative advantages in 2000 were mainly medium-low-tech, labor-intensive industries. Correspondingly, the comparative advantages of China's patent applications around 2000 were also concentrated in medium-low tech areas. However, in 2019, China's comparative advantage shifted to the high-tech field, mainly to the computer and communications industry, and we note that there was a similar shift in the RCA index of patent applications.

Given the strong correlation between knowledge innovation and the comparative advantage in trade, we believe market opportunities can guide China's knowledge innovation activity. However, it also imposes constraints on the leapfrog development of China's knowledge innovation. At present, China does not have comparative advantages in a number of fields, such as aerospace and pharmaceuticals, but these fields are of great significance to China’s development. It is difficult to incentivize knowledge innovation activity in these fields relying solely on market forces. Therefore, how to overcome restrictions from comparative advantages and let the government play its role has become a crucial issue that needs to be addressed in China.

1.2.2.5 Pattern 5: Growth in International Collaboration in Knowledge Innovation Has Slowed in China

Cooperation between scientific researchers from different countries can facilitate the generation of new ideas. It can not only improve efficiency in problem solving, but also provides an important way for sharing the achievements of knowledge innovation. In practice, international cooperation can help researchers cut research expenses, share research resources, and investigate global academic topics.⁴⁶ Thus, it has become an important way to improve the efficiency of scientific research. For example, in 2016, China-US co-authored STEM papers accounted for 23% of all internationally co-authored papers in STEM disciplines in the US and 46% of all internationally co-authored papers in STEM disciplines in China,⁴⁷ indicating that China and the US were key partners for each other in academic collaboration in STEM disciplines, and that China has become an important participant in international academic cooperation.

However, the growth of China's participation in international academic cooperation has slowed in recent years. Since 1995, major developed countries have strengthened international academic collaboration, with the proportion of internationally co-authored papers on the rise. In contrast, China’s international academic cooperation could be divided into several stages (Fig. 1.23). From 1995 to about 2007, the growth in output of domestic papers outpaced that of internationally co-authored papers, resulting in a decline in the proportion of internationally co-authored papers. From 2007 to 2016, China started to accelerate its integration into the global academic system and ramp up international collaborative research, pushing a rapid increase in the proportion of collaborative papers. Since 2016, the proportion of international collaborative papers in China has still been lower than that in developed countries, and the upward trend has shown signs of stagnation or even decline, while the proportion of collaborative papers has continued to rise in developed countries.

Fig. 1.23

China's proportion of internationally co-authored papers grew after 2006 and flattened recently. Source Microsoft Academic, lens.org, CICC Global Institute

The number of internationally co-invented PCT patents in China has stagnated since 2013 between 3,000 and 5,000 per year. It has not risen proportionally to the increase in the overall number of PCT patent applications in China, and the share of international co-inventorship has decreased significantly.⁴⁸ The decline in the proportion of patent applications jointly filed by Chinese and foreign inventors reflects the improvement in China’s independent technological innovation capability. However, it could also exert an adverse impact on the transformation of China’s industrial structure. Since 2013, China has started to push the transformation of its industrial structure by changing its extensive growth model and boosting domestic demand. However, China is positioned toward the lower end of the global value chain with limited need for innovation, while developed countries often demonstrate advantages in innovation on the upstream end of the global value chain. Domestic companies usually establish cooperative relationships with foreign partners before a country starts its industrial upgrading. However, when the country begins to move up along the global value chain, its competition with developed countries may intensify, resulting in less willingness to cooperate. Overall, the stagnation of patent cooperation between China and foreign countries indicates a major loss in social efficiency and requires policy intervention, in our view.

1.3 Innovation Economics in Application for Improving Value Chain Security

As we mentioned in Pattern 5 of the previous section, China is usually positioned at the relatively low end of the global value chain, which is consistent with our analysis in the first section, that China faces risks of decentralization. However, analysis of risks from the perspectives of exports and imports only helps identify whether there are horizontal or vertical risks in each industry, and does not inform us on how to improve industry chain security. As discussed in Sect. 1.1, from the perspective of international competition, the key to improving a country’s industry chain security is to increase the country's irreplaceability in the industry chain. How is irreplaceability measured? Generally speaking, a segment that is more irreplaceable should have stronger pricing power. Among micro-level financial indicators, we believe gross margin can better measure pricing power than ROE. This is because ROE depends not only on the irreplaceability of the company itself, but also on the asset turnover ratio influenced by the business cycle, the net margin influenced by the accounting system, and the leverage ratio that represents the capital structure. Relatively speaking, gross margin can better reflect the company's own pricing power.

Hereinafter, we will use gross margin as a measure of irreplaceability and R&D intensity as a measure of innovation capability to depict the industry chain and R&D chain in an attempt to discuss how to improve industry chain security. It should be noted that not all industries with disadvantages or risks require high policy attention. For example, China's share of global imports is 22% for cosmetics and 6% for pharmaceuticals, and the import source CR4 is 78% for cosmetics and 54% for pharmaceuticals. However, this does not mean that from a policy perspective, more attention should be paid to cosmetics than to pharmaceuticals. Considering the different social significance of each industry, policy makers should concentrate resources on improving industry chain security in key areas.

Based on the above analysis of advantages and disadvantages of various industries, and taking into account the criteria of “systemic importance” and “promising development prospects,” we identify three key areas of industry chain security, namely digital technologies, green industries, and biotechnologies. We believe the digital economy is the most “systemically important” field of China's economy in the next 10 years, and that industrial digitization will bring profound changes to many industries.⁴⁹ The green economy should bring both new constraints and new opportunities to China's economy in the next 40 years under the goal of carbon neutrality,⁵⁰ and we see enormous development prospects for new technologies and industries such as hydrogen energy, carbon capture, and energy storage. The COVID-19 pandemic has highlighted the systemic importance of the bioeconomy. Against the background of global warming, the bioeconomy may become increasingly important for the sustainable development of human society, whether in terms of food production or disease prevention and treatment. The great success of artificial intelligence (AI) in predicting protein structures also means that the development of the digital economy may accelerate the development of the bioeconomy.

1.3.1 Implications from Three Major Areas: Industry Chain and R&D Chain Are Highly Positively Correlated

1.3.1.1 Digital Economy: The US Occupies High-Margin Segments of the Industry Chain Through High R&D Investment

Our analysis of the industry chain and R&D chain shows that China's software industry needs to improve its security. In the software industry, infrastructure software has the highest gross margin and the highest level of irreplaceability. The US is far ahead of other economies in terms of gross margin and revenue of infrastructure software, supported by its much higher R&D investment in this segment. The gross margin of application software is between the gross margins of infrastructure software and IT services. The US also leads in gross margin and revenue of application software, backed by its much higher R&D investment than other economies (Fig. 1.24). IT services has the lowest gross margin and the lowest level of irreplaceability. Although the US has higher revenue in IT services than other economies, it does not have much higher gross margin and R&D intensity in this segment.

Fig. 1.24

Software industry value chain and R&D chain (2020). Note We selected 2,712 companies in the Bloomberg software and technical service sector in Western Europe, the US, China (including Hong Kong SAR and the Taiwan region of China), Japan, and South Korea for calculation. The size of the bubble represents the scale of revenue. Source Bloomberg, CICC Research Software Team

1.3.1.2 Green Economy: R&D Investment Supports China’s Current Leading Position

The green economy is an emerging field in which China currently holds international competitive advantages, and solar and electric vehicle (EV) batteries are two industries in which China has more prominent advantages. In 2019, in the photosensitive semiconductor industry, China accounted for 21% of global exports, much higher than its 7% share of global imports. In the storage battery industry, China accounted for 28% of global exports, more than three times its 8% share of global imports; and its share of global lithium-ion battery (LIB) exports was 38%. An important reason for China's advantages is that the country has maintained a high R&D expense ratio in major segments of the industry, which has enabled Chinese companies to take large market shares and enjoy high gross margins. In the solar industry chain, the midstream and upstream manufacturing segments are dominated by Chinese companies, and the US and Germany only maintain some competitiveness in the inverter segment. It is worth noting that the advantages of the US and Germany in the inverter segment are still based on their high R&D intensity. In the EV battery value chain, Chinese companies have invested more in R&D in various segments and also enjoy higher gross margins than foreign companies (Fig. 1.25).

Fig. 1.25

Solar industry value chain and R&D chain (2020). Note Calculations are based on 2–3 leading companies in various segments of the solar industry in key countries and regions in 2020. Wind Info, Bloomberg, CICC Research Electrical Equipment and Utilities Team

China's current dominant positions in the solar and EV battery industries mean that the country faces different value chain risks in these two industries from those in the digital economy. In the short term, China mainly faces export-related horizontal risk in the green economy. Some developed countries use anti-dumping and countervailing measures to impose sanctions on China's solar and other green industries. In the long run, the green economy should continue to thrive and the more critical technologies for achieving carbon neutrality are energy storage, hydrogen energy, and carbon capture. In these areas, which are of strategic significance to the green economy, China has not yet established advantages similar to those in solar and EV battery industries. Major countries are all investing in R&D in these strategic areas. If other countries make breakthroughs first, China's advantages in the green economy could be significantly weakened, creating new vertical risks.

1.3.1.3 Bioeconomy: The US and Europe Are Highly Irreplaceable in the Seed and Pharmaceutical Industries

Generally speaking, the vertical risks faced by China in the bioeconomy are not as serious as those in the digital economy. China's import dependence on bioeconomy is concentrated in certain segments, such as soybeans and seeds in agriculture, and human vaccines and medical instruments in the pharmaceutical sector. Judging from the pharmaceutical value chain and innovation chain, areas with higher gross margins and irreplaceability largely have higher R&D intensity (Fig. 1.26). China has much lower R&D intensity than the US and Europe in the pharmaceutical manufacturing and medical device industries, which feature higher gross margins. In 2020, the top 10 US pharmaceutical companies by revenue invested 22.1% of revenue in R&D, while this proportion was only 14.3% for the top 10 Chinese pharmaceutical companies. In the agricultural sector, apart from fertilizers which are more like chemicals, other segments largely demonstrate a positive correlation between gross margin and R&D intensity. The segment with the highest gross margin and irreplaceability is seed production. China's R&D expense ratio in upstream seed production is only about 5%, much lower than the levels of 10–20% in the US and Europe. As a result, China has low gross margin and irreplaceability in this segment.

Fig. 1.26

Pharmaceutical industry value chain and R&D chain (2020). Note The size of the bubble represents revenue. The sample is the top 10 companies in the industry in the region. The Chinese pharmaceutical industry is ranked according to the pharmaceutical revenue of listed companies. As a reference, the top 10 companies in the US pharmaceutical industry have revenue of US$28.68bn. Source Wind Info, Bloomberg, CICC Research Pharmaceutical Team

1.3.2 Innovation Economics for Improving Industry Chain Security

To sum up, the fundamental reason for industry chain risks caused by international competition lies in China's insufficient R&D investment and weak technological innovation capability. In this regard, we conducted a detailed analysis in the second section of this chapter to examine and compare China's intellectual innovation capability from multiple perspectives, including academic publications and invention patent applications.

There are two basic ways to accelerate technological innovation in China. One is to absorb technological spillover from advanced countries, and the other is to rely on indigenous innovation. On one hand, certain countries have implemented many restrictive measures against China in high-tech industries; on the other hand, these countries also need China to be integrated into the global industry chain to provide more efficient production. Over the past few decades, China has relied on both absorption of foreign technology and indigenous innovation, which complement each other.

Against the current background of international competition, China is facing an increasingly serious technological blockade, making it increasingly difficult to absorb advanced technological achievements from abroad. In fact, we believe international competition is determined by the objective strengths of countries, and is inevitable. In particular, technological competition holds the key to competition between major powers.

Although the US's expanding technology controls against China have indeed had an adverse effect on China's absorption of advanced foreign technology, this also forces Chinese companies to accelerate indigenous innovation, in our view. An empirical study shows that a China-US technology decoupling would lead to an increase in the patent output of Chinese companies in corresponding areas.⁵¹ This shows that the foreign technology blockade is not only an “industry chain risk”, but is also creating opportunities for spontaneous acceleration of China's scientific and technological progress.

Even if there is no external pressure from international competition, we believe China must strengthen R&D investment and accelerate technological innovation in order to achieve its medium- and long-term growth goals. The amended Communist Party of China (CPC) Constitution in the 19th National Congress of the CPC pointed out: In the new century and new era, the strategic goal of economic and social development is to create a “moderately prosperous society” in all respects by the centenary of the CPC (founded in 1921), and to build China into a modern socialist country that is “prosperous, strong, democratic, culturally advanced, and harmonious” by the centenary of the People's Republic of China (founded in 1949). In the 14th Five-Year Plan (2021–2025) for National Economic and Social Development and the Long-Range Objectives Through the Year 2035, the goal is for per capita GDP to reach the level of moderately developed countries.

The above goals indicate that China’s per capita GDP growth rate will need to be maintained at a relatively high level between 4% and 6% for a long period in the future. However, China’s demographic dividend is fading. We expect China’s working-age population will experience negative growth in the next 30 years. At the same time, since China’s young and middle-aged population has crossed the growth inflection point in 2011 and is in a downward trend, the future savings rate is likely to continue to decline, leading to a gradual slowdown in capital accumulation. This means that in order to achieve China's medium- and long-term development goals, China’s economic growth will rely more on total factor productivity (TFP) growth driven by technological progress.

At the same time, according to a study by Bravo and Marín (2011),⁵² for middle- and high-income countries, every 1ppt increase in the intensity of R&D investment increases the growth rate of TFP by approximately 0.63ppt.

It should be noted that because different researchers have different understandings of relevant data and calculation methods, it is difficult for them to obtain uncontroversial conclusions for quantitative analysis of TFP and even economic growth. Therefore, our calculation results should be better understood as a quantitative analysis to express a qualitative conclusion, that is: In the current demographic situation, increasing the growth rate of TFP is a necessary condition for achieving mid- and long-term growth goals. Therefore, we believe it is necessary to increase the intensity of R&D investment. However, the more urgent question is how to achieve higher intensity of R&D investment and how to accelerate technological progress. We believe policy intervention is the key to these questions. As Nobel laureate Paul Krugman proposed in his new trade theory, for high-tech industries, the government needs to take appropriate intervention measures to foster innovation and gain long-term competitive advantages.

As for what specific policies should be adopted, this is a question to be answered by innovation economics. We conduct a systematic analysis of China’s current state and challenges of technological innovation in Chapter 1 (including R&D resources and talent from the supply side and knowledge innovation achievements from the output side of innovation). Chap. 2 emphasizes the advantages that large scale brings about. China enjoys scale effect derived from the demand side of innovation, namely domestic consumption and international trade. The analysis of China’s innovation economics is only the beginning, and is intended to guide our exploration of specific paths to accelerating technological innovation at the industrial level.

Chapters 3–6 constitute the industry-centered part of this report, mainly focusing on the application of innovation economics in industries of the real economy. On one hand, China needs to rely on technological innovation to improve its irreplaceability in key areas, mainly including digital economy (Chap. 3), green economy (Chap. 4), and bioeconomy (Chap. 5). It should be noted that since different industries have different positions along the value chain and technological characteristics, the focus of innovation economics may vary in each chapter. Domestic demand has played a key role in the rise of China's solar and EV battery industries, but whether it remains important for future strategic technologies in fields such as energy storage, hydrogen energy, and carbon capture needs further analysis. For pharmaceuticals, at a time when China's commercial medical insurance system is not yet mature, whether the domestic industry can leverage its demand generated by foreign commercial insurance to accelerate its progress is an issue worthy of discussion. For agriculture, there is an urgent need to strengthen the protection of intellectual property rights and increase R&D intensity.

On the other hand, it is necessary to discuss manufacturing and logistics industries in particular. As mentioned above, the industrial basis for the formation of division of labor in the international industry chain is the transformation of manufacturing to the Wintel model and the sharp decline in logistics costs. China's large and complete manufacturing sector and well-developed logistics facilities are important advantages at the industrial level for the country to accelerate technological innovation. However, these two important pillars of innovation are “big but not strong.” How should we view the roles of China's manufacturing and logistics industries in supporting innovation? What are the shortcomings to be overcome? What new opportunities do these two industries have to further strengthen their roles in supporting innovation? We will discuss these issues in Chap. 6.

Due to the positive externalities of innovation, the private sector is often less willing to support innovation, thus the public sector plays a major role in correcting market failures. Chapter 7 discusses how the government could build a national innovation system through coordinating and supporting technological innovation. The innovation system of a country includes not only market-based cooperation and interaction between enterprises, universities, and government, but also various innovation-related framework conditions such as infrastructure, policy framework, and macroeconomic environment. Regional innovation centers represent the main components of a country's innovation system. The rise of such innovation centers brings about concentration of innovation resources, and thus can boost the growth of local economies, create innovation-friendly environments, and facilitate the completion of a country's innovation targets. The chapter also stresses the importance of financing innovation. A common view is that external financial support is needed for innovation. The problem is that both real economy capital and financial capital are profit-seeking. If real economy capital is unwilling to provide financing for innovation, why is external financial capital willing to support innovation? In fact, according to our analysis of the relationship between finance and innovation in Chap. 7, financing does not spontaneously go to innovation, and financial intervention does not spontaneously promote innovation. Therefore, the government needs to contribute to innovation by financing through direct investment, system building, and credit enhancement.