The Study Of City Labor Sorting Effects Based On Social Network

In the process of the promotion of China’s new urbanization,labor agglomeration is one of the important way to improve the urban productivity,and the spatial distribution characteristic of labor force is well worth studying.In the traditional research,the explanation of urban productivity source is mainly based on Marshall’s externality theory of new economic geography.With the development of "new" new economic geography theory,the spatial selection effect of heterogeneous individuals is estimated to be another important mechanism of city productivity promotion,thus will overestimates agglomeration effect on regional productivity under the traditional agglomeration perspective.In order to deeply study the source of urban productivity,this paper puts "agglomeration effect" and "selection effect"into a unified framework,namely "sorting effect",to explore the internal mechanism and macro effect of urban labor force.However,the agglomeration and selection of labor force in cities are inevitably affected by the behaviors of surrounding individuals,that is,the social network,which has been ignored in the study of urban economics.Therefore,based on the perspective of social network,this paper will explore the question of labor sorting effect under the influence of social network,including:Firstly,based on the connotation of social network,the micro mechanism of labor sorting under social network is pointed out,including the information bridge function and peer effect of social network,which influence the wage income of labor force,and the rent-seeking function and peer effect,which influence the employment choice behavior of labor force.Then the closed network,based on the strong interaction of kinship and geography,and the open network,based on the weak interaction of work and community,are defined,and difference between the two types of networks is compared.Furthermore,the structural equation model is used to identify the open network and closed network of urban labor force,which is the key to examine the mechanism and macro effect of labor sorting in later content.Specifically,based on the subjective and objective dimensions,the indexes of two types of social network is set up,then use the path coefficient to calculate the open and closed networks index of China’s urban labor,and find that the closed network being inversely proportional to the size of city,open network is proportional to the size of city.Secondly,according to the micro-mechanism of social network,we construct the theoretical model of the urban labor sorting effect under the social network,which provides the theoretical basis and measurement method for the empirical test of sorting effect.The generation process of sorting effect is divided into two stages.The closed network influences the transfer cost of labor force in the urban selection stage,and the open network influences the equilibrium wage of labor market in the city.Based on the formation mechanism,the urban labor force selection model and agglomeration model are constructed,and the urban labor sorting model under the social network is derived by using the nested forms of the two models.Therefore,the theoretical hypothesis of this paper is put forward,that is,closed network brings the selection effect of urban labor force,urban scale forms the homogeneity aggregation effect of labor force,and open network brings the heterogeneity aggregation effect of urban labor force.Then,we empirically test the driving mechanism of the formation of urban labor sorting under the social network.For the wage income decision model,the two stage fixed effects and instrumental variable estimation results confirmed that the open network can significantly promote the improvement of individual wage income,and the results of sub-samples shows that groups with high and middle education degree can obtain salary promotion effect under open network,while the sample of low education degree do not show obvious positive effect.For the employment choice model,the estimation results of Probit model and instrumental variable method show that closed-network can significantly improve the probability of individuals using the relationship in employment.And the results of sub-samples with different education degree show some differences,that is closed network will significantly influence the employment selection with the high and low education degree groups,while the effect on medium education degree group is not obvious.Further,the results of double linear model regression including city and individual show that the micro-effect of open-network is influenced by a variety factors of city level,including the city’s economic development,the process of industrialization,infrastructure and the foreign direct investment and so on,but closed network micro-effect is only affected by the level of urban infrastructure.Finally,the macro-effect of urban labor sorting under the social network is empirically examined by using the method of unconditional distribution and parameter corresponding.By using the Kernel density profile of labor skill,the distribution characteristics of labor skills under different social network strengths are compared and analyzed.Then,the empirical results of sorting effect parameters under different social network strengths is estimated.The results show that:for the different open network groups,the skill advantages of the bigger city are more benefited from homogeneity and heterogeneity agglomeration effects;for the different closed network groups,the sorting of the middle-small city shows positive agglomeration effects when compared to the big city,while the sorting of the middle-small city shows opposite results when compared to the super and huge city.In addition,the analysis results of contribution degree method show that the two kinds of agglomeration effects have similar contribution to skill difference between cities,while the contribution degree of choice effect is worser.The sub-sample estimation results with different education degrees show that the regression results of lower education were more robust,and the regression results of local labor force were more significantly than those of migrant labor.