Log in to access the full article.
In 1937, the Japanese government accelerated the expansion of its military expenditure and began to impose controls on the economy to maintain the balance of international payments. The controls were developed through trial and error. The cotton spinning industry was one of the industries most deeply affected by these controls. Initially, the government simply reduced the allocation of foreign exchange for raw cotton imports. However, because this measure prevented the export of cotton products, especially to countries outside the yen bloc, a new scheme of control, the ‘export–import link system’, was adopted from the second half of 1938. This scheme was intentionally designed to give firms incentives to export to non-yen bloc countries and to incorporate elements of market mechanisms into the system of economic controls. Analysing firm-level data, we find that under the link system, firms with higher labour productivity (LP) tended to grow faster, as occurs under a market economy. This relationship was not observed during the early stage of the controls. This difference is reflected in the pattern of change in aggregate LP. Under the export–import link system, the positive reallocation effect was substantial, similar to a market economy, whereas it was almost zero under the early controls. These findings indicate that the design of controls matters for the performance of controlled economies.