Twenty Years, 500 Interviews: A Peking University Scholar's View of China's Automotive Independence

By Lili Yu

2023 was a year when China's auto industry was examined as a whole. That year, building on its 2022 surpassing of Germany, China overtook Japan to become the world's largest car exporter. But where did China's indigenous auto brands actually begin? For many, this remains hazy. In the national memory, China had its independently produced Hongqi sedan as early as the 1950s, while after the 1980s, the market was essentially flooded with the "old three" and "new three" — products of Sino-foreign joint ventures born from the "market for technology" policy. How China's indigenous auto brands struggled out from the cracks after 2001, and how they moved step by step from imitation to independence, from being crushed by joint venture brands to confronting them head-on, and even achieving reverse export in the new energy vehicle era — this is precisely what Feng Kaidong's interview attempts to reveal.

Feng Kaidong is a scholar at Peking University's School of Government Management who has long studied Chinese industry and innovation policy. As early as 2004, Feng and Professor Lu Feng (author of Light Change: A Company and Its Industrial History) participated as researchers in policy discussions on the development of China's automotive, communications, and other industries. In particular, a research report they submitted to the Ministry of Science and Technology on the "market for technology" policy played a role in promoting independent innovation in China's auto industry.

Over two decades, Feng interviewed roughly 500 person-times of engineers, corporate managers, workers, scholars, and government officials. These stories were collected in The Tide Rises: The Birth of China's Innovative Enterprises, published last year.

Tracing this automotive history matters not only because China's auto industry accomplished a small-scale breakthrough from a "chokepoint" situation, but also because it opens the black box of manufacturing production and organizational processes. In Feng's view, if China's economy is to achieve an innovation-driven transformation and refuses to become a dependent economy, it must open this black box of manufacturing organization and production processes, understanding "how to accomplish upgrading at the organizational or industrial coordination level."

In the independent innovation of China's auto industry, there was not only technological dependence and competition between China and foreign countries, but also inertia inherited from the planned economy era. Behind both threads lies essentially a process of "de-dependence."

Moreover, we can see how SOEs or joint ventures with greater resources, capital, and policy advantages lost to indigenous brands in competition. This is precisely what countless innovation stories have revealed: incumbents will always miss the chance to disrupt themselves due to vested interests or the inertia of past success, and this is where innovative forces find their opportunity.

The conversation follows:

Behind the Lane Change

"Dark Surge": From the early days of using "market for technology" to bring in foreign investment, to becoming the world's largest car exporter by the end of 2023, China's auto industry has completed a new round of role reversal. How do you view the reasons behind this transformation?

Feng Kaidong: In the short term, the Russia-Ukraine war, the COVID-19 pandemic, and shifts in China-US relations were important factors in the explosive growth of China's auto exports. Especially with traditional automotive powers like the US, Europe, Japan, and South Korea — most automakers from these countries withdrew from the Russian and Belarusian markets, giving Chinese auto companies opportunities, notably reflected in the growth of traditional fuel vehicle exports.

Of course, this also reflects Chinese auto companies' evolution from marginal players that mainly performed assembly, manufacturing, and sales within multinational corporations' global systems, to independent players that autonomously control product technology platforms and can cultivate local supply chains from scale and technology.

"Dark Surge": Many people believe that a key reason Chinese cars could move from marginal to independent players was the possibility of overtaking on a curve, because new energy vehicles' "three electric" technologies — motor, electronic control, and battery — replaced the engine and transmission technologies of the fuel vehicle era.

Feng Kaidong: Many people think China's success in new energy vehicles was due to changed technological barriers, but why was China the only one to effectively seize this opportunity? Clearly it wasn't simply that barriers became lower, but that we effectively completed accumulation and recombined resources, then produced effective industrial organization.

"Dark Surge": Which accumulations in China's auto industry do you think were particularly critical in the transition to new energy vehicles?

Feng Kaidong: The "independent innovation" in China's auto industry brought structural transformation. In complex industrial systems, technologies inherently exist in hierarchies: some technologies matter more than others. Whether specific components (technologies) are adopted, and in what form, depends largely on the will of core manufacturers who hold product definition power. If local industry doesn't develop autonomous product technology platforms, the entire industrial system can hardly develop.

Moreover, the auto industry is scale-oriented. For suppliers to develop design capabilities, they need host manufacturers to provide opportunities for trial-and-error磨合, and need complete vehicle manufacturers to provide large-scale orders or expectations to offset their development costs. This required Chinese automakers to blaze a trail in complete vehicle product development, thereby gradually building up local supply system development capabilities.

"Dark Surge": As early as 2004, you and Lu Feng became involved in research on independent innovation in China's auto industry. What were your findings at the time?

Feng Kaidong: Around 2000, the Ministry of Science and Technology had already realized that something seemed off with the previous "market for technology" strategy in the auto industry, and commissioned many researchers to investigate. Professor Lu Feng and I were among them.

One finding was that in the fuel vehicle era, Sino-foreign joint ventures under "market for technology" did not lead to growth in local industrial technological capabilities. Because the practical activities conducted under market for technology were basically production and assembly activities. No matter how much wealth you accumulated in production and assembly, as long as you didn't allocate resources to product development activities and engage in corresponding developmental practices, you couldn't form experience accumulation and capability growth. Later, in a policy analysis report on China's independent intellectual property auto industry that we submitted to the Ministry of Science and Technology, we gave three examples of "viable" industrial innovation: Hafei, Chery, and Geely.

"Dark Surge": Didn't the market for technology agreements include technology transfer provisions?

Feng Kaidong: Many joint venture agreements explicitly required joint development of new products and teaching of corresponding technologies, but foreign parties, leveraging their superior technological position, used many tricks to obstruct our formation of technological capabilities through reverse learning. For example, technical drawings given to Chinese partners often had holes in them, with much information cut from the drawings (such as technical data for some KD parts, etc.).

In the market for technology process, what we controlled was how to transform production factors into products; what we lost was substantive control over production activities. At the time, few developmental activities existed before delivery of product technical drawings in joint ventures. Many technology centers established by contract in joint ventures were essentially ornamental institutions, with little direct connection to technology and product industrialization development.

"Dark Surge": Was this problem never recognized?

Feng Kaidong: People generally assumed at the time that obtaining foreign advanced product drawings for localized production could effectively help us understand how to do better in development and design. But this was impossible. To form genuine capabilities, we needed to trial-and-error ourselves, to interact with upstream and downstream.

Before taking the market for technology path, we had wanted to import foreign equipment on a large scale, but due to insufficient foreign exchange payment capacity and capital shortages at Chinese enterprises, we had to start bringing in foreign investment. The foreign exchange balance issue then pushed policy toward localization of imported product parts. At the time, we mainly wanted to use CKD (complete knock-down) and SKD (semi-knock-down) methods to have more industrial activities occur locally, gradually building industrial capabilities by increasing localization rates.

"Dark Surge": But as conditions improved, why did independent car manufacturing remain slow to happen?

Feng Kaidong: For some SOEs in the auto industry, the earliest pursuit of localization led to organizational resources tilting toward joint venture assembly operations. At the time, FAW's Changchun Automotive Technology Research Center had many technical and product development personnel who jumped to FAW-Volkswagen's joint venture assembly operations because the pay and benefits were better there. Resource tilting often brings shifts in organizational internal discourse power; over time, this forms a kind of mental inertia, manifested as cultural values within the organization about what is "reasonable" and "scientific." Once this rigidity forms, it's very hard for the enterprise to turn back.

"Dark Surge": Would frontline engineers be content with merely assembling rather than completely building a car?

Feng Kaidong: Many engineers were not content. In 1992, FAW had a model called Sankoule, built by several old Hongqi sedan engineers who used scrap materials during their spare time when others were "playing cards and watching TV" to piece together two cars. Later they drove one from Changchun to Beijing. This was particularly symbolic — after all, when the first Hongqi sedan was produced in 1958 (then called "Dongfeng"), it was driven from Changchun to Beijing by then-FAW director Rao Bin, entering Zhongnanhai to receive Chairman Mao's "audience."

This event generated good social response, but because FAW was then busy with Jetta localization with Volkswagen, with resources tilting toward the joint venture, some managers criticized these car-builders as not attending to proper duties, not focusing their energy on the joint venture model.

"Dark Surge": Looking back, this was a very tragic moment.

Feng Kaidong: In 2002, Beijing Automotive built Beijing Hyundai through a joint venture with South Korea's Hyundai to construct local automotive industry in Beijing. This was a huge shock to China's auto industry. Because our automotive industry started in 1953, our sedan industry roughly in 1957-1958 — we already had our own products then — while South Korea's auto industry only truly began with the 1973 "Korean National Car Plan."

For the older generation of Chinese auto engineers, this result was unacceptable. They asked: our engineers' capabilities are not inferior, our history is longer, so what problems led to our cars now being inferior to South Korea's? Something must have gone wrong somewhere, so many people began searching for whether there existed new, potentially successful industrial experiences.

The Battle for Dominance

"Dark Surge": So in the early days, China's indigenous auto brands gathered a group of people who wouldn't accept this?

Feng Kaidong: The earliest people to pursue automotive "independent innovation" very much resembled a "counter-movement" by those people and enterprises not incorporated into "market for technology," seeking to realize their self-worth. Like the two engineers behind FAW's Sankoule model mentioned earlier, who after retirement eventually went to Geely, where Li Shufu entrusted them with important responsibilities. Chery's company address in Wuhu is No. 8 Changchun Road precisely because many engineering and technical personnel came from FAW — they refused to accept defeat and wanted to do it again. Twenty to thirty young Chinese engineers trained in the Dongfeng Citroën joint venture project, upon returning to find their enterprise's departments all doing joint ventures, went to Chery to realize their ambitions. The former director of Tianjin Gear Factory, because his enterprise had joint ventured with Toyota, went to Geely to develop automatic transmissions. These people were also strategically visionary and willing to take risks. Like when Li Shufu was getting into cars, many mocked him — a private enterprise making cars, isn't that suicide? He said, then please give me a chance to jump. During Chery's "grass hut" startup period, Yin Tongyao wanted to convince two university classmates to join; the classmates worried Yin had "fallen for some pyramid scheme," until in the grass hut they lifted a tarp to reveal an oil clay model of a car...

"Dark Surge": If China didn't form its own technological capabilities during the market for technology process, where did later technological capabilities come from?

Feng Kaidong: In terms of specific growth paths for technological capabilities, most indigenous brands early on defined products through reverse engineering, with fairly strong imitation traces. Later, they caught some special windows. Like in the late 1980s to 1990s, many auto groups faced collapse from overexpansion, causing many specialized technical companies spun off from large enterprises, as well as engineering technology enterprises and design studios originally serving them, to have to seek opportunities in underdeveloped markets. Hafei Auto, for instance, encountered Italian design firm Pininfarina, which had designed models for Ferrari, Maserati, and Rolls-Royce. Of course, behind "market for technology," China's WTO accession and integration into multinational corporations' global production networks made the original planned system hard to sustain, and substantive changes in industry management systems also provided a more relaxed environment for independent innovation enterprises to rise from marginal markets.

"Dark Surge": Since market for technology didn't bring technology, what was different about Hafei's cooperation with Italian design firm Pininfarina that you mentioned?

Feng Kaidong: What differentiated Hafei from joint venture automakers or SOEs in the previous "market for technology" was that they very forcefully demanded control of corresponding design processes and dominance. For Pininfarina, such cooperative development was common; they often treated it merely as a turnkey project. Early on their attitude was very arrogant — when giving Hafei engineers a lecture on automotive design, they even started from "what is red, what is blue." This attitude thoroughly angered Hafei engineers. The forcefulness behind this was actually Hafei's sense of urgency for survival. At the time Hafei had roughly 20,000-30,000 people; their aircraft business had shrunk due to reduced defense investment, so they had to find ways for the several thousand people in autos to support the entire enterprise through developing the auto business. For Hafei Auto, this cooperation had no retreat.

"Dark Surge": After gaining dominance, how were technological capabilities formed?

Feng Kaidong: In Hafei's subsequent series of cooperations with Pininfarina, Pininfarina not only conducted body design and provided much process support; most importantly, Hafei formed a workflow for how to complete complex product development through foreign cooperation, including how to combine design and development processes with manufacturing production to form a production system. Of course, this process wasn't accomplished overnight. Hafei Auto only blazed this trail through sustained cooperation with Pininfarina to develop multiple mini-vans; by the time they developed the sedan "Lubao," this path was basically established.

"Dark Surge": What did this workflow model mean for independent innovation in China's auto industry?

Feng Kaidong: In early Chinese indigenous car-making, because there were no drawings, products were basically defined through reverse engineering. Like Chery's Fengyun — it didn't imitate Volkswagen or the Jetta, but rather a Volkswagen-affiliated company's model in Spain; after that company was acquired by Volkswagen, it shared many components with the Jetta, making the two models look very similar. But later when Chery developed its second batch of models like the Eastar and QQ, it basically used a workflow model similar to Hafei's: self-defining systems, breaking down tasks, and conducting extensive cooperation with internal and external teams; authorizing frontline engineer teams to master international mainstream development processes and related technologies through extensive foreign technical cooperation projects; and through a series of product development projects, forming understanding of system integration.

"Dark Surge": How do you view the启示 brought by the Hafei case? What's more critical behind design processes and specific methods?

Feng Kaidong: Design processes and specific working methods were never the big problem. The biggest problems were actually路线问题 and organizational problems — where power and resources tilt. Another key was producing an "engineer-led" organizational model. Technological capabilities cannot be bought; when conducting new product and complex technology development, you can only authorize and rely on frontline engineering and technical teams facing concrete situations, and through resource mobilization, form capabilities in solving specific engineering and technical problems.

"Dark Surge": Looking back at this process, at the time those joint ventures actually held absolute advantages in technology, capital, and policy. How did China's indigenous auto brands step by step reclaim the market?

Feng Kaidong: The starting point for joint ventures then was to develop China's potential market scale, reusing their existing model assets and equipment assets. China had the "old three" (Santana, Jetta, and Fukang); with these products they could survive very well. The German head of Shanghai Santana once said, why consider new product development? We'll just make the Santana model, keep it unchanged for 10 years, then reduce the car's cost to $5,000 per unit, and it will become the world's most competitive car — this logic seems absurd today. Around 2001, after independent innovation enterprises successively obtained car manufacturing qualifications, they not only launched more cost-effective low-price models, but also continuously launched large numbers of new models to develop the market during their technology learning process, while rapidly catching up to joint ventures in product technical configuration and quality. This also forced multinational corporations to later put more new models into the Chinese market; some enterprises that couldn't adapt to this transformation had to leave the Chinese market.

"Dark Surge": In this catching-up war, was the indigenous brands' low-price strategy their most critical weapon?

Feng Kaidong: More precisely, they used low prices to bring into the market portions of people not defined by multinational corporations as consumers. These independent innovation enterprises from after 2000 to 2010 were basically making large-scale investments in new product development, frequently launching new products. They provided higher cost-performance by controlling labor costs, increasing localization ratios, even tolerating lower gross margins, thereby creating new markets. At the time, the cheapest Jetta among the "old three" still cost around 100,000 yuan, while Chery QQ could sell for just over 30,000. Chery's SUV products, including the Tiggo that later ignited the entire SUV market, were roughly 80,000-100,000, while Toyota and Honda equivalents were around 200,000. Japanese companies disassembled Hafei's "Songhuajiang," "Zhongyi," and "Minyi" mini-vans; GM also disassembled Chery QQ — their conclusion was that at equivalent costs, they couldn't build such cars at the time.

The Rise of Individual Enterprises and the Transformation of a System

"Dark Surge": Manufacturing involves very long production chains. Did independent innovation enterprises encounter particular difficulties in joining the existing order?

Feng Kaidong: Why did some parts in Chery's early models bear a certain German brand's logo? Cars are complex large-scale industry; facing a completely new enterprise, supplier enterprises couldn't be sure whether making molds for them was worthwhile — making molds for just one enterprise carried risks. Chery used some personal connections to persuade certain suppliers to provide parts; to make parts interchangeable, they had to adopt some components from that German brand's system in product design. When making the Fengyun model, one of Chery's founding veterans, carrying drawings to visit suppliers, found that no connections worked — because no one could say whether Chery would survive, and these suppliers were also wary of multinational corporations' attitudes. This founding veteran recalled that the initial start was very difficult, hitting walls everywhere; he cried all the way while carrying the drawings.

"Dark Surge": Returning to the present, after China-US relations trended toward tension in 2017, the original globalized industrial chain is also being restructured. What new changes do you think this entirely new situation will bring to China's independent innovation?

Feng Kaidong: Around 2005, the independent innovation transformation in China's auto industry was actually the rise of individual enterprises themselves. It emphasized that local enterprises could dominate complex technology and product development. After 2017, what China faces is actually the systemic transformation of entire industrial chains or innovation chains.

"Dark Surge": How to understand systemic transformation?

Feng Kaidong: Most independent innovation enterprises have very strong internal vertical integration characteristics. Look at Huawei making submarine cables, inverters, LiDAR — they do everything themselves because it's a bamboo shoot poking through, with no suitable companies to solve supporting problems, so they have to do everything themselves. Chery developed its own machine tools, its own forklifts, its own robots; even Wuhu Shipyard (an SOE from one of the 156 projects during the First Five-Year Plan) is now Chery's. They exported complete vehicles to the Third World very early, but because exporting complete vehicles using general container ships had very high disassembly costs, to produce dedicated ships, Chery had to merge the shipyard. Previously Chinese enterprises were all on global chains, so domestic systemic relationships were actually very weak. Before 2017-2018, Huawei wouldn't give its product OEM to SMIC because it had to compete with Apple, so it needed to cooperate with the world's best foundry TSMC. SMIC before 2017-18 basically had no relationship with SMEE because it would only purchase the world's best lithography machines, so SMEE previously had no opportunity to form corresponding capabilities through supplying and interacting with SMIC.

"Dark Surge": But if everything is limited to independent innovation and internal circulation, won't it fall into closed innovation?

Feng Kaidong: As early as 2005 when discussing independent innovation, some people also understood it as closed innovation. The fact is, China's auto industry independent innovation was very open from the start, and it grew entirely within the great melting pot of international cooperation as China joined the WTO and integrated into globalization. The "independent" here actually emphasizes Chinese industry's mastery of dominance over product and technology development, is "de-dependence," not "closed innovation," and certainly not "building cars behind closed doors."

"Dark Surge": When emphasizing independence, many people feel this concerns an internal-external logic of openness versus closure, but listening to you, what you want to emphasize more is a vertical logic of dominance versus dependence.

Feng Kaidong: Essentially, independent innovation emphasizes the logic of market competition. I often tell people that the market for technology model actually has great similarity to the planned economy model, because both lead to local enterprises not mastering autonomy over technological activities. During the planned economy period, R&D work was generally given to research institutes, design was executed by design institutes, then production was done by state-owned factories. FAW's older generation of engineers once told me that FAW's not-so-good independent development capability couldn't entirely be blamed on them. Because in the 1950s, FAW's startup construction was modeled after the Soviet Union's Stalin Auto Plant. And the Stalin Auto Plant itself had no R&D center, only a product technology section, whose configuration and functions were mainly to preserve drawings rather than develop new products, because Soviet auto plants long depended on inputs of American, German, and later Czechoslovak model platforms. Later, the Changchun Automotive Technology Research Center was merged into FAW to become FAW's technology research center, but after engaging in Sino-foreign joint ventures, besides early Hongqi and commercial vehicles, most production activities fell back into localized manufacturing production models, causing the technology research center's strength to not be fully applied to developing local passenger vehicles.

"Dark Surge": Japan and South Korea's auto industries both went through processes of catching up and surpassing Western powers. Did they also fall into similar states of dependence and de-dependence?

Feng Kaidong: Japan never entered a state of dependence. They did fall into debates between developmentalism and comparative advantage theory. At the time, MITI (Ministry of International Trade and Industry) played an important role in this debate; later developmentalism gained the upper hand, enabling them to unify thinking very early after the war — that is, to develop Japanese industry, especially the auto industry, through developing frontier technologies and developing world markets. South Korea was more dramatic, providing two different paths. Their auto industry originated in the 1973 National Car Plan. In this process, South Korea's two largest auto enterprises Hyundai and Daewoo diverged in their paths. Hyundai was selected to participate in the National Car Plan; they broke away from CKD assembly patterns similar to Sino-foreign joint ventures and took the OEM route. From integration to self-design, until the 1990s, they achieved autonomous forward development. Daewoo, meanwhile, continued depending on the American enterprise system for assembly, though later nominally developed some of its own models. The final result was that Daewoo directly went bankrupt and was acquired by GM during the 1997-98 financial crisis. Hyundai, after one failure, successfully entered the US market in the 1990s. Their promise of 5-year/100,000-kilometer warranty when entering the US market remains an international quality standard to this day.

"Dark Surge": Why is research on industrial history and innovation so scarce in China?

Feng Kaidong: Globally, under the influence of mainstream economics, industrial history research has undergone a process of marginalization. Attention to market transactions has overwhelmed attention to production processes; the strengthening of finance and capital has replaced former emphasis on production. Domestic teaching also draws from mainstream economics — not studying concrete industrial organization problems, not studying practical technical problems, differences between enterprises and enterprises are highly abstracted, as if this is more "scientific." Yet innovation and innovation competition are themselves processes where a minority of innovators squeeze old market forces and open new spaces through "creative destruction." Intellectual circles today actually seriously lag behind manufacturing development practice. If we want to study manufacturing development, in the final analysis we must still understand its development mechanisms from organizational, political science, and other levels.

"Dark Surge": You've said that compared to The Tide Rises, you hope more to have the opportunity to write a book about personal stories in the independent innovation process. Which people or stories particularly moved you?

Feng Kaidong: Too many. For example, in early 1990s Northeast old SOEs, when enterprises encountered construction capital shortages, in order to build new factory buildings, from cadres to ordinary workers, through all-staff meetings unanimously agreeing, everyone used their spare time to do volunteer labor building the new factory. In 1999, when Taiwan had a major earthquake, a FAW retired old engineer representing Chery at Taiwan Fuzhen Mold Company supervising mold development, at the first moment of the earthquake, without hesitation threw himself onto the mold, using his body to protect the enterprise's "hope for independently developed models"...

"Dark Surge": In the opening section of the book, you mentioned that your turn to Chinese industrial policy research came from a survey at a Foshan foundry. Seeing those assembly line workers at the time, what truly pierced you?

Feng Kaidong: What pierced me at the time was actually empathy. Before university, I had never left my hometown; university was my first time leaving the southern border province where my hometown was. I did fairly well in school, surviving relatively high selection ratios all the way from junior high to high school, then from high school to university (Tsinghua University Department of Mechanical Engineering). But in this process, you see countless peers from your youth, even close friends around you, falling behind — perhaps many of your classmates went to work after junior high, or started families and businesses after high school, making a living in small trades. In this process, you realize that the 2,000 people in the factory, whether theoretically or practically, could all be another you.

Image source | IC photo

Layout | Yunxiao Guo