The sudden outbreak of COVID 19 brings another reason, national safety, and diversification, to move factories out of China. A more complex multi-centered system of global production web would definitely bring additional stability, but will also be associated with additional cost. For an economist to find out whether the relocation is a good idea in the sense of diversification, we need to estimate the relationship between the marginal increase of social welfare from additional stability and the cost, which would be a very hard job.
Here I would take a step back. I would argue that the relocation of manufactures would not be easy even if the marginal increase in welfare is larger than the cost, because the manufacturing of products involves extensive coordination along production chains. If these coordination problems were not settled, relocation would mostly require continued subsidies or high tariffs (similar to the Import Substitution Industrialization from the 1950s to 1970s in developing countries).
Modern manufacturing requires specialization to enhance efficiency. For instance, a typical passenger automobile is made up of about 20,000 parts, and the final assembler produces only 20% of them (measured in value-added). Therefore, if a country does not have the ability to produce upstream parts and components, setting up a factory of final assembling (Completely Knock Down, CKD) would lead to low profits. Besides, the continued demand for importing parts and components would bring pressure to the balance of payments. When China sets up the joint venture of ShanghaiAuto-Volkswagen in 1985, the cost of importing all the parts of a Santana to Shanghai is even higher than importing an assembled vehicle. Without localization of component production, the manufacturing of automobiles in Shanghai would never be economically viable without government subsidies or high tariffs.
However, the localization of upstream manufacturers is also a hard job. For the case of Volkswagen Shanghai, after three years of CKD production, the localization rate of Santana is only 2.7% (tires, horns, radio antennas, and labels). The low speed of upstream localization can also be observed in other countries, for instance, the TV set industry in Mexico in the 1980s and the semiconductor industry in Japan in the 1970s.
Localizing upstream suppliers is hard because it involves the coordination and bargaining among firms. For upstream firms, they face high uncertainty because they need to invest a lot in machine tools, personnel training, and trials, pass numerous tests, and then negotiate the contract with downstream firms. For instance, Volkswagen Shanghai required all new suppliers to pass the test in its German factory and then a road-trial of 60,000km. The huge uncertainty and the long time-span associated with this process, together with the high initial investment, deterred entrepreneurs from entering the supply chains.
Downstream firms usually lack the incentive to promote supply chain localization if they enjoy the subsidies from the government. Since they have stable suppliers from abroad, switching to local suppliers would only happen when the locally produced parts are in good quality, but the prices are sufficiently low. Even if the downstream firm really wants to cultivate local suppliers, achieving a high level of localization would be hard because the coordination process mentioned above needs to repeat hundreds of times. The downstream firm would need to, for each component, identify the best candidate local supplier, help it enhance product quality and production scale (may even need to help secure the funding of investment in some cases), negotiate an interest-compatible contract, and conduct continuous monitoring and assessment before mutual trust is established. Completing this process requires long-term vision, patience, good local connections, and the ability to shoulder the costs. The fact that few firms demonstrate these qualities may explain why localizing manufacturing in a globalized economy if hard.
I am definitely not suggesting that it is impossible to relocate manufacturing. However, the above analysis shows that simply paying for moving costs is not enough to attract factories. There needs to be a way to mitigate or to overcome the high cost of production chain coordination. In the case of China, the government stepped up to provide the needed coordination effort. In 1987, to promote the localization of Santana, Shanghai municipal government set up the Santana Localization Association, participated by Volkswagen, and many car component companies (they used to be suppliers of the outdated passenger vehicle called Model Shanghai). With the strong support of both local and the central government, this association was in charge of helping each of the suppliers tackle problems in meeting the quality requirement of Volkswagen, such as contacting foreign machine tool suppliers, obtaining foreign exchange loans, and training personnel. In 4 years, the localization rate of Santana increased from 2.7% to 83.3%, and more than 180 suppliers obtained the recognition of Volkswagen, and later becomes suppliers of other car-makers, such as BMW and Audi.
By no means am I suggesting that government interference is omnipotent, but economic theories highlights that in some scenarios, a trustworthy coordinator is needed to overcome coordination failure. Depending on the context of different societies, this coordinator can be a large corporation, a social organization, or a government. The essence is that it coordinates the activities of other agents through both economic and noneconomic channels. Otherwise, relocation would probably remain in twitters.
