关键词: 森林公园, 旅游效率, 影响因素, 数据包络分析, 面板数据回归模型

Abstract: 【Objective】 In this study, 31 provincial level unit's forest parks in Chinese mainland were taken as research objects, to study tourism efficiencies of forest parks of mainland China, and provide reference for efficiency promotion of Chinese forest parks. 【Method】 Firstly, based on the DEA computing technique, with tourism income and tourists number as output variables, and the forest parks area, the same year fund investment and staff population as input variables, the input-oriented constant scale model in DEAP2.0 edition software was used to survey the tourism efficiencies of various provinces forest parks of each year from 2009 to 2013. All data came from the Chinese National Forestry Bureau official website announcement. Secondly, based on the supply-demand theory, the population density, the urbanization rate, GDP per capita, the high rank tourism resources density, the forest park density, and the ClassⅠ-Ⅳ highway density of the province, as well as the Woodland forest coverage rate, the fund Investment intensity and the Labor intensity of the forest parks of the province were chosen as independent variables, and the tourism efficiency was the dependent variable, to establish panel-data multi-dimensional linear regression model, and quantitatively analyze factors of influencing forest park tourism efficiency. In this study, all independent variable data were obtained from the Chinese statistics yearbook, the dependent variable data came from the first step computing results. 【Result】 The results show that: 1) In the whole, the mainland forest park tourism efficiencies increase year by year since 2011, but the forest park tourism efficiencies exist big difference among various provincial capital areas. The forest park tourism efficiencies of some provinces undulate greatly during various years. 2) The stata12.0 software was used sequentially to carry on the individual effect examination, the time effect examination, the hausman examination and the different variance steady examination of the regression analysis model, showing that the model exists individual effect and the time effect, and the effect of the stochastic effects model is better than that of the fixed effect model, thus the bidirectional random effect model was chosen as the final regression model (5% significant level). 3) The robust GLS estimation shows that, the population density and Woodland forest coverage rate (5% significant level), the urbanization rate (10% significant level) have the significant positive influence to the forest park tourism efficiency; GDP per capita and the park investment intensity have the significant negative influence to the forest park tourism efficiency (5% significant level); The tourism resources, the forest park density, ClassⅠ-Ⅳ highway density have the positive but not significant influence to the forest park tourism efficiency (10% significant level); The labor intensity has negative but not significant influence to the forest park tourism efficiency. 【Conclusion】 Finally further studies problems are proposed: during 2009—2013, tourism utilization efficiency of Chinese forest parks had significant improved, but there existed large efficiency differences between some provinces, and the efficiency of forest parks in some provinces fluctuated greatly each year. Population density, urbanization rate, tourism resource level, forest park density, traffic development level have played positive roles on the efficiency of forest parks, and the investment density has played a significantly negative impact on the efficiency of forest parks, which is contrary to people's knowledge about the role of capital investment in forest parks. Paper finally gives some advices on how to promote the efficiency of Chinese forest parks, and points out some problems which need further follow-up research.

Key words: forest park, tourism efficiency, influence factor, data envelop analysis method, panel data regression model