Developing economy to get rid of poverty is the common goal of all low-income countries. For many developing countries, small-scale markets and inadequate incentives for private investment are key reasons for economic stagnation. Therefore, these governments often invest directly in the industrial sectors of specific regions or industries, expecting that such industrial policies can promote industrialization and accelerate economic growth. Considering the scope economic characteristics of the industrial sector, the industrial policy of governmental direct investment can make the economy jump from low-level equilibrium to high-level equilibrium. In practice, the policy effect also depends on a series of constraints faced by a country, and these constraints are endogenous to the past economic structure, which means that quantitative research based on the historical perspective is of great significance. It helps to identify the key elements for development and provides empirical support for policy implications.

The difficulty in identifying the long-term effect of industrial policies is that the areas of industrial investment are not randomly selected. In order to achieve causal inference, this paper exploits the long-term effect of the quasi-natural experiment of Anti-Japanese War on the economic development of the rear area. By sorting out historical data files, the county-level industrial database of ten provinces during the Anti-Japanese War is constructed, and the data are matched with the contemporary economic and social development data in the same geographical space. Then, using the regression model, the long-term performance of industrial investment in history is estimated. The main findings are as follows: First, the more the number of industrial enterprises in the county from 1942 to 1945, the higher the level of economic development in the new China period, but the effect is gradually weakening with time. Second, the industrial investment in the rear area of modern times has a positive effect on the long-term economic development through path dependence and initial material capital accumulation. In history, the higher the industrial investment, the higher the degree of industrialization, but there is no significant difference in the production efficiency of industrial enterprises. Third, historically, areas with more industrial investment can attract more people to move in and have a higher urbanization rate, but this historical basis does not significantly affect rural related variables, such as infant mortality rate and net income of rural residents.

The long-term research based on historical events provides a new perspective for the formulation of industrial policies. The case of Anti-Japanese War shows that industrial investment against comparative advantage will cause efficiency loss in the short term, but may bring a positive effect in the long term. This raises a new intertemporal trade-off problem for policymakers about the present and the future, whose subjective time discount rate is the key to solving the problem. In addition, the quantitative history study also reminds us that one should fully recognize the initial conditions of regional development before policy implementation, which means that one can neither copy the experience of other countries nor simply reverse what policies are needed from development goals. Instead, one should be more aware of what policies are really feasible under current constraints.