关键词: 地理距离, 方言距离, 劳动力流动, 引力模型, 大数据

Abstract: This paper studies the impacts of geographical distances and dialect difference on spatial mobility of labor in China in a systematic way by using Baidu big data. Based on a gravity model and the ordinary least square method (OLS), it is found that both geographical distance and dialect difference inhibit the labor mobility through construction of ratio of mobility opportunity. In terms of spatial distance, labor prefers to move to nearby cities. The ratio of labor mobility opportunity decreases by about 0.6% if the geographic distance increases by 1%. In terms of spatial location, labor is inclined to move to the areas in similar dialects. The ratio of labor mobility opportunity falls by around 2% if dialect difference is added by 1%. To alleviate the potential endogeneity, by constructing the gaps of rainfalls and wheat planting suitability between two cities chosen as instrumental variables for dialect difference, two stage least square method (2SLS) shows the result is robust. Furthermore, taking into account the standard Mandarin, the inhibition effect by dialect difference is still robust while the high popularity of the standard Mandarin plays a significant gravity role in attracting labor mobility in the destination. It proposes that popularizing the standard Mandarin, expanding traffic construction and promoting secondary education be the options for policy makers.

Key words: Geographical Distance, Dialect Difference, Labor Mobility, Gravity Model, Big Data