改革开放40年来中国经济持续高速增长，伴随着经济结构的变化，继续以资本和要素投入推动经济发展的经济增长方式已经不可持续。教育和人力资本发展是推动经济发展转型的关键。因此，如何通过教育发展推动全要素生产率增长，提升经济发展的效率，是一个需要加强理论支撑和实证研究的现实问题。关于我国经济发展讨论最多的“拐点”主要有两个，其一是“刘易斯”拐点，其二是人口红利拐点。刘易斯拐点与我国目前所处的二元结构相关，是刘易斯对发展中国家的经济增长特征进行观察所提出的概念。我国的“人口红利”优势渐渐式微，劳动力短缺以及人口老龄化逐渐成为制约经济增长的关键因素。无论是刘易斯拐点还是人口红利拐点，都将着眼点放在人口或劳动力的数量和年龄结构变化上。随着经济的增长，我国人口和劳动力的质量也在不断提高。如何准确衡量劳动力质量对我国经济增长的贡献？教育人力资本将如何影响我国的长期经济增长水平？本文试图在对我国教育人力资本进行长期估计的基础上对上述问题做出回答。

第一，基于中国人口普查数据（1990/2000/2010）运用人口学方法重建了中国1990-2017年历年分年龄、性别、受教育水平和城乡维度的劳动力人数，在此基础上，对我国历年劳动力人口的受教育程度、性别以及城乡差异进行分析，对分级教育和经济增长趋势进行分析。第二，考虑到中国教育发展还反映为地区发展水平的异质性特征，运用分位数回归对各级教育在经济增长的不同分位点处影响力度的异质性；第三，考虑到教育发展与经济增长在时间维度上潜在的双向因果关系，运用动态面板模型研究各个省份教育发展对经济增长的动态影响效应。第四，考虑到中国教育发展与经济增长在空间维度上的巨大差异以及彼此的空间关联性，使用推算得到的我国受高等教育劳动力比例和初中等教育劳动力人口比例进一步分析人力资本对我国经济增长影响的时空效应分析，运用空间杜宾模型纳入空间因素研究教育发展影响经济增长的空间溢出机制。主要得到以下结论：

（1）我国的平均受教育年限整体呈显著上升趋势，全国从1990年的6.82年上升到了2017年的10.36年，农村从1990年的6.04年上升到了2017年的8.68年，城镇从1990年的8.78年上升到了2017年的11.49年。全国高中及以上受教育程度人口占比从1990年的0.02上升到了2017年的0.18，农村的占比从1990年的0.002上升到了2017年的0.05，城镇的占比从1990年的0.066上升到了2017年的0.27.

（2）各级教育规模在经济增长的不同分位点处影响力度出现了明显的差异，对初等教育而言，较高分位点处的参数估计值随着分位点的增加减小了很多，说明对经济发展较为落后的地区，提高教育程度有更迫切的需求；对中等和高等教育而言，经济发展快的地区对接受中等和高等教育的人才需求越强烈。

（3）在空间层面，教育发展的空间关联主要表现为经济上的相互联系，单

纯地理上的空间关联并不显著。不同空间权重矩阵下初等、中等和高等教育规模估计值均在1%的显著性水平上为正，表明各级教育能够显著地促进本地区经济增长，同时空间溢出效应估计值亦显著为正，表明教育水平存在空间溢出效应，邻近地区的教育发展规模和水平对本地区的经济增长有正向促进作用。从效应分解来看，高等教育规模的直接效应系数显著为正，说明高等教育规模能够显著促进区域内经济增长，即高等教育规模水平每提高1%，区域内经济增长将提升7.7%。高级人力资本间接效应为0.168，即高等教育规模水平每提高1%将带来邻近地区经济增长提升16.8%，具有显著的空间协同效应。邻地高等教育规模效应的系数显著高于本地，表明高等教育规模对邻地的经济增长影响更突出，邻地效应占总效应的一半以上，进一步证明了提升人力资本水平过程中由要素资源流动带来的空间外溢效应对经济增长有重要的促进作用。

        Over the past 40 years of reform and opening up, China's economy has continued to grow at a high speed. With the changes in the economic structure, the economic growth mode that continues to use capital and factor inputs to promote economic development is no longer sustainable.Education and human capital development become the key to the transformation of economic development.Therefore, how to promote the growth of total factor productivity through education development and improve the efficiency of economic development is a practical problem that needs to strengthen theoretical support and empirical research.There are two main "inflection points" about China's economic development, one is the "Lewis" inflection point, and the other is the demographic dividend inflection point.Lewis' inflection point is related to the current dual structure of China, and it is a concept proposed by Lewis to observe the characteristics of economic growth in developing countries.As China's "demographic dividend" advantage gradually weakens, labor shortage and population aging have gradually become the key factors restricting economic growth.Whether it is the Lewis turning point or the turning point of the demographic dividend, the focus is on changes in the number and age structure of the population or labor force.With the economic growth, the quality of China's population and labor force is constantly improving.How to accurately measure the contribution of labor quality to China's economic growth?How will the education human capital affect China's long-term economic growth?This article attempts to answer the above questions on the basis of a long-term estimation of China's education human capital.

        Firstly,Based on the Chinese census data (1990/2000/2010), demographic methods were used to reconstruct the number of labor force in China by age, gender, education level, and urban-rural dimension during the 1990-2017 calendar year.On this basis, this paper analyzes the education level, gender, and urban-rural differences of the labor force in China over the years, and analyzes the trend of graded education and economic growth.Secondly, considering that the development of education in China is also reflected in the heterogeneous characteristics of regional development levels, using quantile regression to the heterogeneity of the impact of various levels of education at different quantiles of economic growth.Thirdly, considering the potential two-way causal relationship between education development and economic growth in the time dimension, a dynamic panel model is used to study the dynamic impact of education development on economic growth in each province.Fourthly, considering the huge differences in the spatial dimensions of China’s education development and economic growth and their spatial relevance,using the estimated proportion of labor force in China with higher education and proportion of labor force with primary and secondary education analyzes the space-time effect of human capital on China's economic growth.And using the spatial Dubin model to incorporate spatial factors studies the spatial spillover mechanism of education development affecting economic growth.The main conclusions are as follows:

(1) The average number of years of schooling in China has shown a significant upward trend. The country rose from 6.82 in 1990 to 10.36 in 2017; rural areas rose from 6.04 in 1990 to 8.68 in 2017; cities and towns rose from 8.78 in 1990 to 11.49 in 2017. The proportion of the population with high education level and above in the country increased from 2% in 1990 to 18% in 2017; the proportion in rural areas increased from 0.2% in 1990 to 5% in 2017; the proportion in urban areas increased from 1990 6.6% rose to 27% in 2017.

(2) There is a clear difference in the impact of the scale of education at different quantiles of economic growth at all levels. For elementary education, the parameter estimates at higher quantiles decline a lot as the quantile increases，explaining that there is a more urgent need to improve education in areas with relatively backward economic development.And for elementary and higher education, the regions with fast economic development have stronger demand for talents who have received secondary and higher education.

(3) At the spatial level, the spatial correlation of educational development is mainly manifested in economic interconnections and the purely geographical spatial correlation is not significant.The estimates of the scale of elementary, secondary and higher education under different spatial weighting matrices are positive at the 1% significance level.It shows that all levels of education can significantly promote the economic growth of the region.At the same time, the estimated value of the spatial spillover effect is also significantly positive, indicating that there is a spatial spillover effect in the level of education, and the scale and level of education development in neighboring areas have a positive role in promoting economic growth in the region.From the perspective of effect decomposition, the direct effect coefficient of the scale of higher education is significantly positive, indicating that the scale of higher education can significantly promote the economic growth in the region, that is, for every 1% increase in the level of higher education, the economic growth in the region will increase by 7.7%.The indirect effect of advanced human capital is 0.168, that is, every 1% increase in the scale of higher education will bring about 16.8% increase in economic growth in neighboring regions, with a significant spatial synergy effect.The coefficient of the scale effect of higher education in neighboring areas is significantly higher than that of the local area, indicating that the scale of higher education has a more prominent impact on the economic growth of neighboring areas.The neighborhood effect accounts for more than half of the total effect, which further proves that the spatial spillover effect caused by the flow of factor resources in the process of increasing human capital has an important role in promoting economic growth.