Abstract: Abstract：Based on the DEA-Malmquist model and provincial panel dataset(2005-2016), this paper measured the provincial forestry total factor productivity(TFP) in China, and identified the spatial spillover relationships between provinces by the VAR Granger causality test. On this basis, in the framework of social network analysis, the structural characteristics of spatial correlation network and its effects are analyzed. The results showed that, the spatial spillover effect of forestry TFP was commonly existed in China, and there was multiple superposition in the inter-provincial spatial forestry TFP spillovers;the correlation network presented a trend of multi-core development, and the trans-provincial efficiency spillover showed an obvious gradient feature; the gap between provinces contributed to the expansion of forestry market′s boundaries, besides, similar innovation level and financial development level can promote technology exchange and cooperation between provinces,and the “Matthew effect” of forestry growth. Therefore, more attention should be paid to the spatial disequilibrium pattern of China′s forestry production efficiency. Through macro-control and market system construction, favorable conditions should be created for the cross-regional flow and efficient allocation of production factors, and forestry development policies should be formulated and implemented according to local conditions to give play to the functions and advantages of each region.

Key words: forestry total factor productivity, spatial network, social network analysis, DEA-Malmquist model, VAR Granger causality test