Abstract: Under the new normal of economy, it is necessary to improve the efficiency of the allocation of industrial capital and other factors to realize the transformation of economy to high-quality growth.Under the government led and investment driven industrial development mode, combined with the threshold variables of local debt pressure and financial development difference, the paper uses the dynamic panel smooth transformation regression model to test the gradual evolution relationship between local debt scale and industrial capital allocation efficiency in a non-linear framework.It is found that the appropriate scale of local debt can improve the efficiency of capital allocation, but with the increase of debt pressure, the positive effect of debt on capital allocation gradually weakens and produces negative effect.The reason is that moderate debt can make up for the short board of industrial infrastructure, reduce the cost of land, produce leverage effect on industrial development, and thus improve the efficiency of capital allocation; however, excessive debt can promote financial risk, occupy credit resources and strengthen corporate financing constraints, resulting in low efficiency of capital allocation.In addition, the study found that the scale of local credit can ease the financing constraints of borrowing and promote the capital flow among industrial industries; and the deepening degree of industrial finance can enhance the role of market competition mechanism in factor allocation, alleviate the investment mismatch caused by borrowing, so as to enhance the efficiency of capital allocation. The above analysis results show that local governments should be more rational, moderate and standardized in borrowing.

Key words: local government debt, regional financial differences, industrial capital allocation efficiency, panel smooth transformation regression model