Wei Chen and Shiping Tang argue that industrial policy succeeds when a powerful lead agency can coordinate procurement, technology transfer, firm discipline and research support.
This is a very good article and I agree with its core conclusions (I'm honored to see they cite some of my research, too). I would respectfully push back somewhat on Chen and Tang's characterization of auto sector industrial policy as a "dismal failure". I think this is broadly true for the 1994-2004 period for the reasons they emphasize and more, but in my view they somewhat understate the direct and indirect contributions of industrial and technology transfer policies post-WTO to the sector's subsequent development. I also think it's important to remember that creating globally-competitive conventional auto brands is extraordinarily difficult. There's a reason only one country -- South Korea -- has managed to do so post-World War II that did not have a robust auto manufacturing sector before the war, and it took South Korea two-plus decades to get there. There's no doubt industrial and technology transfer policies in conventional autos underperformed relative to leaders' (unrealistic) expectations, but I think ten or even twenty years is too short a time frame to assess these policies' full impact. That said, I do appreciate that this article periodizes the sector's policy trajectory. Too much commentary over-indexes on the 1980s and 1990s and extrapolates into the future. What I would like to see more research on is whether, to what extent, and in what ways conventional auto sector industrial policy-linked development contributed to the subsequent rise of NEVs. My own cursory research suggests the effects were non-trivial, though also not decisive.
This is an excellent article. China’s experience with industrial policy — both its successes and its failures — offers important lessons for many other late-developing countries.
Beyond the conventional automobile case discussed in the article, China’s early CPU programs, several core software initiatives, and its regional aircraft projects all went through long periods of underperformance. The problem was not simply a lack of state investment or a lack of planning. These sectors had longer technology-accumulation cycles, more complex ecosystems, more uncertain market windows, deeper supply-chain bottlenecks, and user demand that was much harder to shape through administrative coordination.
That is why the success or failure of industrial policy cannot be judged merely by whether there are policy documents, subsidies, ministries, planning agencies, or national programs.
The more important questions are:
Is there a sufficiently large market demand?
Can the demand side be organized?
Do domestic firms have the ability to learn and iterate?
Is the supply chain already close to a critical threshold?
Is the technological pathway clear enough?
Can foreign technology be absorbed, rather than merely purchased and assembled?
Can market competition create pressure for upgrading, rather than low-level redundant construction?
Can state institutions provide support while also imposing discipline?
Many countries today want to copy China’s industrial policy. But simply copying planning documents, subsidy tools, industrial parks, investment-promotion slogans, and state-backed funds will usually not be enough. Industrial policy is not a policy menu that can be mechanically transplanted. It has to be embedded in the specific market structure, technological stage, firm capabilities, fiscal capacity, bureaucratic system, and international environment of each industry.
More precisely, successful industrial policy requires at least three layers of conditions.
The first layer is state capacity: the ability to set long-term goals, coordinate ministries, discipline local governments, organize financing, set standards, and enforce implementation.
The second layer is industrial conditions: whether the technology path is catch-up friendly, whether the value chain can be decomposed, whether domestic firms can learn, whether suppliers can grow, and whether the market is large enough to support repeated iteration.
The third layer is market mechanism: whether there is real demand, competitive pressure, profit space, an exit mechanism for failed firms, and enough entrepreneurial incentive to convert policy support into actual product capability.
Without any one of these layers, industrial policy can easily turn into inefficient subsidies, redundant capacity, local protectionism, technological stagnation, or long-term dependence on foreign firms.
This is a very good article and I agree with its core conclusions (I'm honored to see they cite some of my research, too). I would respectfully push back somewhat on Chen and Tang's characterization of auto sector industrial policy as a "dismal failure". I think this is broadly true for the 1994-2004 period for the reasons they emphasize and more, but in my view they somewhat understate the direct and indirect contributions of industrial and technology transfer policies post-WTO to the sector's subsequent development. I also think it's important to remember that creating globally-competitive conventional auto brands is extraordinarily difficult. There's a reason only one country -- South Korea -- has managed to do so post-World War II that did not have a robust auto manufacturing sector before the war, and it took South Korea two-plus decades to get there. There's no doubt industrial and technology transfer policies in conventional autos underperformed relative to leaders' (unrealistic) expectations, but I think ten or even twenty years is too short a time frame to assess these policies' full impact. That said, I do appreciate that this article periodizes the sector's policy trajectory. Too much commentary over-indexes on the 1980s and 1990s and extrapolates into the future. What I would like to see more research on is whether, to what extent, and in what ways conventional auto sector industrial policy-linked development contributed to the subsequent rise of NEVs. My own cursory research suggests the effects were non-trivial, though also not decisive.
China isn't planning a D-Day invasion of Taiwan; they’re executing a slow, systematic strangulation. While the U.S. waits for domestic chip fabs to come online in 2030, Beijing is tightening the noose today using the Russian playbook. Here is why a sliver of silicon is the most dangerous geopolitical flashpoint on earth.https://triggledger.substack.com/p/the-silicon-choke-point-why-the-next?utm_source=share&utm_medium=android&r=8gc1qf
This is an excellent article. China’s experience with industrial policy — both its successes and its failures — offers important lessons for many other late-developing countries.
Beyond the conventional automobile case discussed in the article, China’s early CPU programs, several core software initiatives, and its regional aircraft projects all went through long periods of underperformance. The problem was not simply a lack of state investment or a lack of planning. These sectors had longer technology-accumulation cycles, more complex ecosystems, more uncertain market windows, deeper supply-chain bottlenecks, and user demand that was much harder to shape through administrative coordination.
That is why the success or failure of industrial policy cannot be judged merely by whether there are policy documents, subsidies, ministries, planning agencies, or national programs.
The more important questions are:
Is there a sufficiently large market demand?
Can the demand side be organized?
Do domestic firms have the ability to learn and iterate?
Is the supply chain already close to a critical threshold?
Is the technological pathway clear enough?
Can foreign technology be absorbed, rather than merely purchased and assembled?
Can market competition create pressure for upgrading, rather than low-level redundant construction?
Can state institutions provide support while also imposing discipline?
Many countries today want to copy China’s industrial policy. But simply copying planning documents, subsidy tools, industrial parks, investment-promotion slogans, and state-backed funds will usually not be enough. Industrial policy is not a policy menu that can be mechanically transplanted. It has to be embedded in the specific market structure, technological stage, firm capabilities, fiscal capacity, bureaucratic system, and international environment of each industry.
More precisely, successful industrial policy requires at least three layers of conditions.
The first layer is state capacity: the ability to set long-term goals, coordinate ministries, discipline local governments, organize financing, set standards, and enforce implementation.
The second layer is industrial conditions: whether the technology path is catch-up friendly, whether the value chain can be decomposed, whether domestic firms can learn, whether suppliers can grow, and whether the market is large enough to support repeated iteration.
The third layer is market mechanism: whether there is real demand, competitive pressure, profit space, an exit mechanism for failed firms, and enough entrepreneurial incentive to convert policy support into actual product capability.
Without any one of these layers, industrial policy can easily turn into inefficient subsidies, redundant capacity, local protectionism, technological stagnation, or long-term dependence on foreign firms.