2000—2015年中国高学历人才分布格局及其影响机制

基于2000年、2010年中国人口普查分地级及以上行政区数据和2015年中国各省份1%人口抽样调查数据，采用变异系数、泰尔系数、基尼系数测度中国高学历人才比重分布的空间不均衡程度，并采用面板数据Tobit随机效应模型，识别影响高学历人才比重空间分布的主要因素。结果表明：① 中国的人才比重分布表现出极大的不均衡性，人才比重的高低与城市等级密切相关，主要表现在直辖市、省会城市、计划单列市等行政区吸引了大量人才，而普通地级城市人才比重相对较低；② 10 a间人才比重的变异系数、基尼指数和泰尔指数均有所下降，表明人才比重分布的空间不均衡程度有所下降；③ 平均工资水平、生活成本、城市等级、每万人高校在校学生数、每万人医院卫生院床位数、人均科学事业、教育事业经费与人才比重呈正相关，中学生师比、万人互联网用户数与人才比重呈负相关，失业率、单位面积二氧化硫排放量、绿地率对人才比重没有影响。

Spatial Pattern and Determinants of Highly Educated Talents in China, 2000-2015

Based on the data from the 2000 and 2010 China censuses by prefecture level and above, and the 2015 1% population sample survey data of Chinese provinces, this article studies the spatial pattern evolution of the distribution of highly educated talents and its influence mechanism in China. The coefficient of variation, Theil coefficient and Gini coefficient are used to measure the spatial imbalance of the highly educated talents proportion. The panel data Tobit model is used to identify the main factors affecting the spatial distribution of the proportion of the highly educated talents. The results show that: 1) The distribution of the highly educated talents proportion in China shows a great imbalance, closely related to the city level, which is highlighted in administrative regions such as municipality, provincial capitals, and city specifically designated in the state plan, to which a large number of talents are attracted, while the talents proportion in ordinary prefecture-level cities is relatively low; 2) In the past 15 years, the coefficient of variation, Gini coefficient and Theil coefficient of the talents proportion have all declined, indicating that the spatial imbalance of the talents proportion has decreased; 3) Average wage level, cost of living, city hierarchy, students enrollment of regular institutions of higher education per 10 000 people, the number of hospital beds per 10 000 people, expenditure for education per capita, expenditure for science and technology per capita are positively correlated with talents proportion; The ratio of middle school students to teachers and the number of internet users per 10 000 people are negatively correlated with talents proportion; Unemployment rate, green rate, sulfur dioxide emissions per unit area have no significant impact on the talents proportion.