Getting China's Industrial Policies Right

Are China’s industrial policies changing? If so, how? Some high-level facts are well known, like China’s ambitions described in the famous Made in China 2025 (MiC2025) plan. But analysts need more granular information to understand whether China’s industrial policies are changing in practice and how that may affect businesses on the ground. Unfortunately, existing resources have not been able to address these questions in a systematic way.

In partnership with Policybot, a policy data provider and market-investment intelligence technology firm, we are developing a data-driven approach to track China’s industrial policies and help investors and corporate managers make better sense of them.

To understand the changes of policies, one must look at historical data to detect policy patterns and contextualize them in China’s political and economic environment. To assess the impact of policies on businesses, one needs to look at a specific location or sector where policies could be very different from the macro picture. In short, high-quality evidence and analysis should be the foundation of good decision making.

In the sections below, we are excited to present you our initial works.

At the national level, we discovered a shift of policy focus from green development to innovation capacity in the past year. But at provincial level, we found that Guangdong still pays more attention to green development than innovation, leaving significant room for firms to explore their own approach of technology advancement. For a specific industry, robotics, we found that policies in Guangdong have been creating demand and building infrastructure to facilitate development of the industry rather than directly promoting the industry itself.

These findings reveal interesting nuances buried in high-level analysis of China’s industrial policies which are actually critical for businesses to make the right decisions at the right time. Transparency of provincial governments is improving, and we offer solutions to capture the trend. We are also developing new capacity to track more ministries and news, identifying new dimensions of policy influences such as subsides and regulations, and correlating policy announcements with market dynamics.

We started by exploring industrial policies at national level with a focus on the MiC2025 plan. We trained the machine to collect all policies published on the websites of relevant authorities, including the National Development and Reform Commission (NDRC), the Ministry of Industry and Information Technology (MIIT), and the Ministry of Commerce (MOFCOM). The machine found 475 policies framed as part of the MiC2025 plan.

Our search confirmed that policies stopped referring to MiC2025 in July 2018 (Figure 1), around the same time when domestic media stopped doing so. Mentioning of the MiC2025 plan was at its peak level in the second half of 2016, suggesting that it took authorities about a year to translate the grand plan into implementing policies. MIIT was expectedly the most important actor in issuing relevant policies, and NDRC also caught up in early 2017. In contrary, MOFCOM, which oversees China’s trade and investment with foreign entities, was mostly silent throughout the period, suggesting that MiC2025 was mainly targeting domestic audience.

A more interesting question is whether China has adjusted its industrial policies since 2018. One may guess “yes”, but finding evidence to test that guess is difficult, not to mention identifying the differences. The machine again provides a solution – after reading those 475 policies, it identified a group of key words that are frequently used in MiC2025 policies, which we grouped into four categories based on high level objectives of the plan: innovation capacity, quality improvement, green development, and digitalization. We then illustrate the weight of these objectives in Figure 2 – data suggests that policies used to focus on green development, but has shifted significantly towards innovation capacity since October 2018, after US sanction on ZTE made it clear to Chinese leaders that self-reliance of key technology is critical to sustain the country’s development path.

Provincial-level Nuances: the Case of Guangdong

While the central government sets the direction of China’s development, provincial governments still have significant liberty to shape local plans and implementing policies. Having this in mind, we taught the machine to collect and analyze industrial policies in Guangdong province – China’s manufacturing base and growth engine. To avoid noises related to the MiC2025 plan as discussed above, we choose to focus on policies promoting advanced manufacturing, which started in early 2000s, included in various industrial upgrading plans (including MiC2025) afterwards, but were not picked up by headline news well. By feeding the machine with official definition of advanced manufacturing, we found 528 relevant policies from the Guangdong government website.

A closer look into the search results generate three interesting findings. First, government transparency in Guangdong has improved significantly in the past three years. Out of 528 policies, 314 (60%) were published after 2017, whereas only 214 (40%) policies were published in the 15 years before then (Figure 3). This is exciting news for researchers, who have been aware of the differences between central and local policies for a long time but lacked sufficient data to understand them. As China’s economy slows, it is more important than ever to track provincial policies to detect any structural opportunities or early signs of reform. With improved transparency and our joint capacity, such exercise is finally possible.

Second, Guangdong pays more attention to some industries and cities than others. As illustrated in Figure 3, green development (new energy, environment protection, etc.) is the focus of nearly 50% of Guangdong’s industrial policies, while only 20% and 16% policies focus on telecom and smart manufacturing respectively. We also found 302 policies with a specific city focus, which we illustrate in Figure 4. Among them, 103 policies addressed Guangzhou, the capital city of Guangdong, followed by 81 for Shenzhen, China’s innovation hub, 21 for Zhanjiang and 15 for Dongguan.

These results are somewhat surprising – one would expect more policy support in telecom industries which are the key enablers of China’s industrial ambitions. But in fact Guangdong has been paying much more attention to green development. It is possible that green development also depends on some kinds of telecom development therefore the share of policy attention to telecom industries is underestimated in our approach. But it is more likely that the government of Guangdong, a province that benefited the most from China’s market liberalization, is willing to let the market and the firms explore their own path of technology upgrade. Similarly, Shenzhen receives less policy attention than Guangzhou. In other words, telecom industries and Shenzhen are less dependent on policies and therefore more market-driven than many would have expected, an important insight that high-level policy analysis cannot provide.

Finally, most of the policies appear ad-hoc instead of well-planned. By teaching the machine to read through text and pick up mentioning of China’s five-year plans (FYPs), we were able to identify policies following the FYPs. Figure 5 shows that out of 511 policies issued after 2005, only 15% mentioned China’s FYPs, with the lowest percentage found in telecom industries (4%), and highest percentage found in energy (24%). That is consistent with the general understanding that technology-intensive industries are growing so fast that cannot be planned ahead, while energy is one of the most controlled and regulated industries in China. But for all industries, most of the policies were not planned, reflecting provincial governments’ ability to adapt to changing dynamics, a feature that researches focusing on China’s long-term plans would completely miss.

Capture What Matters for a Specific Industry

What matters to businesses most are policies surrounding a specific industry. We hereby narrow down to robotics, a priority industry in China’s industrial plans. After we identified stakeholders of the robotics industry and defined key words for each stakeholder, the machine helped us sort out 58 policies that are directly relevant to the robotics industry in Guangdong, out of 528 policies promoting advanced manufacturing discussed above.

We have proved that at national level policy focus are shifting towards innovation capacity, and at provincial level that Guangdong has been more active and transparent since 2017. But as presented in Figure 6, policies surrounding the robotics industry did not significantly increase until 2019, and only a few addressed robotics specifically. Before then, policies mainly focused on advocating industrial upgrade that is expected to create demand for the application of robotics technology. Since 2016, the government started building infrastructure (telecom, key common technology, etc.) and environment (business environment, industrial parks, etc.) to facilitate development of the industry.

These dynamics are critical for foreign businesses to manage their investments in China: if policies are creating more demand, then foreign companies should be able to enjoy a share of market growth; development of infrastructure and business environment are also beneficial for foreign companies; but foreign companies should be more careful when policies started exercising direct influence over the industry – a sign of increased room for the government to seek rents. For the robotics industry, direct government intervention did not start until the escalation of trade war convinced Chinese leaders to shift policy focus towards innovation. In Guangdong and especially in Shenzhen, however, firms are still relatively less dependent on policies, meaning a better competitive environment for foreign businesses but also healthier Chinese competitors.

Stay tuned for more insights and data from Plenum: